DE102013222211A1 - Device and method for detecting signal interference - Google Patents

Abstract

The invention shows a device for detecting signal interferences with a processor device and a comparison device. The processor means is adapted to calculate a power of a component of a received signal having an exponent greater than 2 and to output a result based on the power. The comparison device is designed to compare the result with a predefined threshold value and to provide a signal interference result as a function of the comparison.

Description

The invention relates to an apparatus and a method for detecting signal interference in the radio-based communication by evaluating the received radio signal.

Radio-based communication is also increasingly used in industrial automation technology, enabling more cost-effective and flexible manufacturing processes and the implementation of new automation concepts. The requirements for radio-based communication, in particular reliability and time behavior, for example real-time capability, are significantly higher for this application than, for example, in the home and office sector.

The fact that all radio-based communication systems share and compete for the limited resource radio spectrum proves to be problematic in meeting this requirement. The competition is particularly in the so-called Industrial, Scientific and Medical (ISM) bands, d. H. Frequency ranges that can be used without allocation by regulatory authorities in industry, science, medicine, domestic and similar areas, large. In addition, high-frequency and low-frequency interference occur, for example, by microwave systems or arc welding or by transformers or motors. It should be noted that the propagation of electromagnetic waves is not localized.

Through the so-called frequency management or coexistence management, which, for example, by the VDI / VDE guideline 2185 is defined, the coexistence of radio applications can be established and guaranteed, which is the prerequisite for meeting the requirements. The problem with this is that frequency management can only guarantee coexistence between detected systems. Third-party systems, such as mobile phones with WLAN, Bluetooth, etc., and interference, such as those mentioned above, can interfere with radio-based communication and thus adversely affect their reliability and timing.

Approaches to mitigate these effects describe, for example, in [1] that the requirements of the frequency management z. B. by "... an automatic system for the permanent monitoring of the radio spectrum to be installed, which may already detect irregularities, even if they have no effect on the plant operation" Can be considered. An implementation of such an "automatic system for the permanent monitoring of the radio spectrum" is described in [2]. The spectrum (power versus frequency) in the 2.4 GHz ISM band is measured and compared with the typical environmental conditions (mask) determined in a learning phase. If the measured power density spectrum exceeds the reference value by an adjustable threshold value, the system triggers an alarm via one or more interfaces.

Since foreign systems or interferers can also occur with a power below the reference value and collisions of systems with equal cutoff frequencies can not be detected on the basis of the spectral analysis, this form of realization is limited in principle to ensure coexistence and thus the availability and the time response.

The list used by Bluetooth systems, for example, is the list of frequencies that are not to be used due to interference (blacklist). This information is used to detect sources of interference based on channel patterns, the course of the fault and the bandwidth of the fault.

[1] also describes a method for estimating a signal-to-noise ratio (SNR), which can provide conclusions about a channel quality.

For example. Bluetooth technology offers affordable and relatively uncomplicated solutions for various applications in the so-called ISM frequency band. In applications that place heavy demands on reliability and real-time capability, such as in industrial automation, it may be critical to discover interference sources that can corrupt data packets, which can lead to retransmissions and delays. In an industrial environment where certain frequency bands are reserved for certain piconets (locally limited personal area network (PAN) communicating over, for example, Bluetooth), radio channels are usually predefined, such as for Bluetooth in 15 is shown.

Bluetooth became very popular due to its low cost and low energy requirements. The benefits of Bluetooth extend over several industrial applications, such as sensors and actuators for controlling moving machine parts, measuring and controlling moving objects, coordination between robots and vehicles, machine monitoring and many others. These applications may require strict reliability and Quality of Service (QoS) standards, otherwise the efficiency in terms of resources used, such as time and / or money, may be compromised, as described in [5]. Consequently, maintaining reliable communication within industrial automation environments can be extremely critical.

Signals are modulated in the Bluetooth standard using Gaussian Minimum Shift Keying (GMSK) with a modulation index of 0.35. A binary 1 is represented by a positive frequency deviation, a binary 0 by a negative frequency deviation. GMSK is a form of Continuous Phase Modulation (CPM) that uses filters with a Gaussian shape for pulse shaping. The transmission power can reach up to 100 mW, which can extend the transmission range up to 100 m. Bluetooth systems have 79 different channels, each with 1 MHz bandwidth and used center frequencies through f = (2402 + k) MHz, k = 0, 1, 2 .... 78 (1) as in [6] and exemplarily in 15 is described.

Radio transmitters are forced to co-exist due to extensive use and an increasing number of devices increasingly scarce, as described in [7]. In order to maintain a certain level of industrial QoS, it is important to recognize mutual interference. Bluetooth-Bluetooth coexistence has been addressed in several papers, such as [8], [9] or [10]. Results of this work conclude that Bluetooth Bluetooth interference depends more on the density of nodes in a given range than on the number of nodes. Willig explains in [11] that there is a higher risk of interference if the same frequencies are temporarily occupied by overlapping piconets (hopped). In [12], Howitt derives an analytical model for characterizing mutual interference in Bluetooth piconets. Further, the likelihood of a data transmission detrimental collision requiring re-transmission is deduced, and several factors, such as the frequency shift and the relative strength of both signals, are taken into account.

In addition to interference from overlapping piconets, interference from external systems and devices in a piconet can also affect transmission quality.

Known approaches can be distinguished in energy detectors, detection by hypothesis tests, information-theoretical detectors and detection by Hidden Markov modulation. Energy detection is one of the most common ways to detect interference due to its low complexity. After a Fast Fourier Transformation (FFT) with N nodes, the total power of the signal is compared as the sum of the powers at the N frequency points and with a predefined threshold. However, this threshold depends largely on the level of noise that will vary in a current operation, as described in [13]. Therefore, overall success decreases in high-noise environments. In addition, such detectors only consider the total power of the received signal. An erroneous decision is made when two signals interfere but their total power is less than the threshold or when a single usable signal has a power greater than the threshold. For example, it is conceivable that two possibly uncorrelated or slightly correlated signals of a useful signal transmitter and a jammer trigger a collision of data packets, but at least partially cancel each other out in amplitude by superposition, i. H. Overlay, extinguish. This can lead to a transmission power which is not or only slightly increased but lies below a power or energy threshold value detectable at the interference detector. In other words, in this case the signal interference at an energy detector can go undetected.

When detected by hypothesis testing, the detection of interference can be seen as a classic detection problem where the detection theory can be used. The radio receiver evaluates the probability of two hypotheses: the presence or absence of interference, which is done for each frequency line or spectral line. The hypothesis probabilities are then compared to a threshold value by a generalized likelihood ratio test, as described, for example, in [14]. However, due to the complexity of Probability Density Function (PDF's) used, an analytical evaluation of the threshold may be practically unworkable. In addition, an evaluation of the channel ratios requires knowledge of the number of interferers in each frequency line so that the complexity will increase exponentially with the number of interferers.

Information-theoretical detectors can estimate the number of interferers in a given frequency line by eigenvalue analysis, as described in [15]. Instead of estimating an energy threshold, a covariance matrix is estimated to generate eigenvalues that serve as an estimate of the spectral power. Subsequently, information-theoretical criteria are used to conclude the signal space dimensions and thus the number of interferers. However, the signal space dimension defines the energy threshold used in energy processes so that such detectors are expected to provide similar results as energy detectors.

In hidden Markov modeling detection, an interference detector is based on Hidden Markov chains, as suggested in [16], for example. A state quantity represents the number of possible interferers in each frequency line, where the first state represents only noise, the second state noise plus one signal, the third state noise plus two signals, and so forth. Thus, the state sequence for the Markov chain determines the interference status of each frequency line. However, determining the optimal model for each situation requires large memory requirements and high computational complexity, as described for example in [16]. In addition, a transition matrix that determines the various transition probabilities between any two states should be determined in advance.

Further references are in
US020050061057A1
US000008295189B2
US000008290100B2
US000007149205B2
WO002008112928A2
US020080192644A1
US020080279093A1
US020080299932A1
JP000S58133060A
JP000S58137337A
JP000H08316996A
JP002002335196A
JP002002353878A
JP002008054303A
JP002009267855A
GB000002451682A
CN000101047486B
CN000101692631B
CN000101697624B
US20090197556A1
US6130907A
US6704378B2
EP2391022A1
specified.

It would therefore be desirable to have a device and a method which, with relatively little computational effort, enable a reliable detection of signal interferences.

The object of the present invention is therefore to provide an apparatus and a method for a more reliable detection of interference in a radio-based communication.

This object is solved by the subject matter of the independent patent claims.

The core idea of the present invention is to have realized that the above object can be achieved by allowing results and evaluations based on powers of signal components with an order greater than 2, conclusions on a signal interference generated by an interfering transmitting node.

According to one exemplary embodiment, a device for detecting signal interferences of a received signal comprises a processor device and a comparison device. The processor device is configured to increase a power of a component of the received signal with an exponent greater than 2 calculate and output a result based on the power. The comparison device is designed to compare the result with a predefined threshold value and to provide a comparison result, for example a detected signal interference, based on the comparison.

An advantage of this embodiment is that a signal interference is independent of a total transmission power, which is possibly radiated by a plurality of interfering transmitter in a frequency band, detectable and an increased accuracy or reliability of the comparison result can be obtained.

Further advantageous embodiments are the subject of the dependent claims. Preferred embodiments of the present invention will be explained below with reference to the accompanying drawings. Show it:

1 a schematic representation of a device for the detection of signal interference;

2 a schematic representation of an apparatus which is another embodiment;

3a a schematic representation of a bandpass filter, which is realized by software and based on a control signal is adjustable;

3b a schematic representation of a device having a configurator and the threshold memory;

4 Excerpts of a device in which the processor means is adapted to detect a predetermined number of samples of a complex valued input signal;

5 a schematic representation of a device that in contrast to the device from 4 has only one signal path;

6 a schematic representation of a device for detecting signal interference, which is designed to form the power in the form of a kurtosis of the real parts and the imaginary parts of the complex input signal;

7 a schematic representation of a device for detecting signal interference, which is designed to form the power in the form of a kurtosis of the real parts of the complex input signal;

8th a schematic representation of a simulative structure for obtaining thresholds;

9 a schematic representation of an experimental setup with a device with laboratory equipment for the determination of thresholds;

10 a qualitative graph on which the abscissa denotes a SINR in dB and on the ordinate a respective determined kurtosis value for the comparison of simulatory and experimental threshold values;

11 a schematic flow diagram of a method for deciding whether a signal can be classified as being interfered, ie faulty or not interfered;

12 a graph with five data series 96a e. The abscissa of the graph plots the interferer-to-noise ratio (INR) in dB. The ordinate shows a probability of misdetection;

13 a graph with three data series 98a -C, which compares possible kurtosis values for varying radio standards with varying SINR values;

14 shows a schematic representation of a system model for describing the received signal;

15 shows a schematic representation of channels, as occupied for example by Bluetooth in the 2.4 GHz ISM band.

Before embodiments of the present invention are explained in more detail in detail with reference to the drawings, it is pointed out that identical, functionally identical or equivalent elements, objects and / or structures in the different figures are provided with the same reference numerals, so that shown in different embodiments Description of these elements is interchangeable or can be applied to each other.

1 shows a schematic representation of a device 10 for the detection of signal interference. contraption 10 includes a processor device 12 , a comparison device 14 and a threshold memory 16 , The processor device 12 is configured to calculate a power based on a component, such as a real part and / or an imaginary part of a complex received signal. The power has an exponent of greater than 2, for example, greater than or equal to 2.5, greater than or equal to 2.8, greater than or equal to 3, greater than or equal to 4, or greater than or equal to 5. The processor device 12 is further configured to obtain a result, for example by performing further computing steps, based on the power. For example, the result may be a value independent of a signal power of the useful signal and / or the interference signal, which makes it possible to detect a signal interference, even if, for example, a cumulative signal power of useful signal and interference signal is below a power limit value, so that an interference detection based on the signal power (power or energy detectors) would detect no signal interference.

Alternatively, it is also conceivable that the processor device 12 designed to compute a fourth-order moment, the kurtosis, another power, or another central moment. For example, a third-order central moment, which can also be called a skewness, and hence a power with an exponent of 3, can be calculated. Alternatively, possibly gaining detection accuracies, a central fifth or higher order moment may also be calculated. Alternatively, it is also conceivable to calculate the power with an exponent in the form of a real number, such as 2.1 or 2.3, 3.5 or 4.2. A second-order central moment corresponds to a signal power minus the mean value of the signal and hence the variance. In other words, the second central moment thus corresponds to an average power value. The central third order moment, for example, may be affected by additive white Gaussian noise. This may possibly lead to a reduced detection accuracy, but may have a reduced computational cost over the kurtosis. For example, if the central third order moment is sufficiently accurate, reduced computational effort and hence reduced number of computational operations can result in time and energy savings.

A power with an exponent greater than 2 can thus enable the information gain of the signal superposition.

The comparison device 14 is designed to be the result of the processor device 12 with a predefined threshold, so that a signal interference detection result can be obtained depending on the comparison. In other words, a comparison result can provide conclusions about a superimposition of a useful signal with an interference signal if the result is greater than or equal to the predefined threshold value. The signal interference result may, for example, be used by another device or device to derive actions based on signal interference result, such as alarm or signaling (warning) for other communication points.

The received signal can, for example, from a receiving device 18 , such as an antenna for receiving a wirelessly transmitted signal, or a cable, such as a coaxial cable, for receiving a wired signal. For example, the wired signal may be wirelessly received by a receiving device located at another position of the processor device via a wired or wired connection between the receiving device and the processor device.

The threshold memory 16 is configured to store the predefined threshold value, such that the predefined threshold value is determined by the comparison device 14 can be obtained to perform the comparison. The threshold memory may have a plurality of predefined threshold values, which are predefined, for example, based on different scenarios and / or different reception, transmission or interference types of the received signal.

Thus, for example, various threshold values for interference of a Bluetooth signal to be received, or of a WLAN signal to be received, can be stored on the threshold value store. alternative or additionally, for example, different thresholds for different scenarios of a communication environment, for example an industrial or warehouse, may be stored on the threshold memory, the threshold values depending on whether the warehouse is full, ie, many reflections of radio signals take place, or empty, ie It may reflect a radio signal that is less often reflected.

contraption 10 makes it possible, inter alia, to detect interference from external sources, that is to say from a respective coexistence management system, such as mobile telephones or other electronic devices or other piconets, which may not be entered in a respective coexistence management concept. In other words, a normalized fourth order cumulant, the so-called kurtosis, can be used to characterize the interference intensity in different interference scenarios.

contraption 10 thus makes it possible to detect the superimposition and thus interference of several transmission signals in any received signal. By virtue of the mathematical procedures described in the following, in contrast to known systems, interference that occurs can not only be suspected, for example if a transmission power threshold value is exceeded, but the interference itself can be detected by identifying the signal overlays.

Will device 10 For example, when used to detect interference in a wireless local area network (WLAN) for interference detection, one may be in the threshold memory 16 stored predefined threshold be lower than in a Bluetooth system. An increased medium bandwidth can lead to a smaller predefined threshold value for the same desired bit error rate. For a WLAN according to IEEE 802.11G For example, channels may have a bandwidth of 20 MHz. In contrast, there is a bandwidth of a Bluetooth device of 1 MHz.

The overall system allows itself to detect interferences with a relatively low spectral power density.

For example, the profile curves between a SINR threshold and the power can be constructed by inputting example signal vectors in software such as MATLAB and calculating the power, eg, a central moment with an order greater than 2, of the vector (s) , as it is, for example, in the 8th is shown. Alternatively, the profile curves can also be determined experimentally, as in the 9 is shown.

Since many mobile radio standards, such as GSM or Bluetooth allow a dynamic adjustment of the transmission power to send, for example, with a good reception quality with a lower transmission power and thus save energy in, for example, mobile devices, for example, to extend device lifetimes, it is conceivable that an overlay two transmission signals, each with low power to an interference or a data collision and thus a loss of a data packet can lead, however, a common transmission power of the two transmission signals is below the predefined power or Energieschwellwertes.

An advantage of this embodiment is that by a power of a component of the received signal with an exponent greater than 2 signal interference detection can be made independent of a power of a desired signal and / or interference, so that a reliable signal interference detection is feasible. Reliable means in this regard that a signal interference, which may, for example, lead to a collision of data packets, can also be detected if, for example, a transmission power of Nutzsignalsender and / or interferers are below a threshold based on a maximum signal power.

In addition to the above-described increase in the reliability of radio systems, applications in the field of cognitive radio (CR) can benefit from this form of interference detection. Cognitive radio is a process whereby a wireless communication device perceives its environment and opportunistically accesses the unused spectrum. In this case, a CR must recognize when a primary user reclaims its licensed frequency band and immediately releases the claimed band without disturbing interference.

Previous CR approaches rely either on a database query (TV White Space Devices) - in which all currently available frequency ranges are deposited - or on Spectrum Sensing, in which the channel is checked before the actual transmission to a primary user. However, using Spectrum Sensing may interfere with the primary user. Reasons for this can be found in so-called hidden terminal problem or the simultaneous or short staggered beginning of the radio transmission of the primary user and CRs.

This can be remedied by creating a network of CRs equipped with the interference detection described above. These allow the other CRs to detect interference during transmission, thereby enabling early release of the channel to the primary user.

Furthermore, applications in medical technology or in the industrial environment, such as in the machine control, which can make high demands on a reliable data transmission feasible.

2 shows a schematic representation of a device 20 which is another embodiment. contraption 20 has the processor device 12 on which is a bandpass filter 22 with a variable bandwidth and a calculator 24 includes. Further, device 20 a configurator 26 on top of that with the processor device 12 or the bandpass filter 22 connected is. The configurator 26 is configurable and configured for a first or a second wireless communication standard, depending on the configuration on the first or the second wireless communication standard, a first predefined threshold or a second predefined threshold, which may differ from the first predefined threshold, to the comparator 14 to deliver.

Different wireless communication standards may be, for example, Bluetooth or Wireless Local Area Network (WLAN) and may provide communication rules that define wireless transmission of signals in different frequency ranges and / or bandwidths. For example, a channel can be Bluetooth standard, as shown schematically in the 15 is set up within the ISM band at about 2.4 GHz in one of 79 channels with a bandwidth of about 1 MHz. Alternatively, WLAN, for example. Through the Standards 802.11b, g or n , Areas of the same frequency band with channels in a bandwidth of, for example, 20, 40 or 80 MHz use. The bandpass filter 22 can also be adjustable with respect to a center frequency, that is, by shifting the center frequency of the filter, a variable sub-frequency band is filtered out of an overall spectrum, for example. In the form of a channel selection of a wireless communication standards.

The configurator 26 is for the first or the second wireless communication standard based on a user input on an input interface 28 or automatically configurable by analysis of the received signal. Thus, for example, an input to an input terminal 28 by a user a configuration of the configurator 26 trigger. Alternatively, the configurator 26 For example, perform an analysis of the received signal for used frequency bands or bandwidths used.

Based on the configuration of the configurator 26 the first or second wireless communication standard is the configurator 26 trained, the bandpass filter 22 to set to a first or a second bandwidth, so that one with the bandpass filter 22 filtered received signal substantially only in frequency ranges in which an interference detection is to be performed, having signal components. The bandpass filter 22 may be implemented in software or hardware, and may be, for example, a variable bandwidth filter or a combination of at least two fixed or variable bandwidth filters. The configurator 26 and the threshold memory 16 can also work in a common device 32 be arranged.

The calculation device 24 is designed to use the powers to be based on the bandpass filter 22 filtered component to calculate the power to obtain the result.

It is advantageous in this embodiment that in a communication environment in which multiple wireless communication standards are used for communication, signal interference may be detectable for multiple or all wireless communication standards used. Thus, for example, a disturber can send out a WLAN data packet to an extent that lies below the predefined for the WLAN standard first threshold value, and the reception of a Bluetooth data packet in a scope that is above a second predefined threshold value by a transmitted interference interferes with interference, so that for the WLAN or the Bluetooth communication, a corresponding comparison result from which, for example, an alarm signal can be derived is obtained.

3a shows a schematic representation of a bandpass filter 22 which is realized by software and based on a control signal, for example. By the configurator, adjustable.

3b shows a schematic representation of the device 32 that the configurator 26 and the threshold memory 16 having. The device 32 is for example designed to be based on user input at the input interface 28 a choice element 34 , for example, a switch to adjust. The choice element 34 is configured to form a signal path with a first element 36a indicating a selection of the first bandwidth and / or a signal path having a second element 36b indicating a selection of the second bandwidth to close. Alternatively or additionally, both signal paths can be closed together if, for example, a third bandwidth is a combination of the first bandwidth and the second bandwidth. The first, second or possibly a third predefined threshold value can be obtained via the signal path and provided to the comparison device.

The configurator or the device 32 may be implemented wholly or partly in software, so that, for example, a signal path is a query of a respective predefined threshold value from a memory module and / or the switch element comprises program code for distinguishing a signal received from the input interface.

4 shows excerpts of a device 40 in which the processor device 12 is configured to detect a predetermined number of samples of a complex-valued input signal. The samples may be, for example, by sampling an analog signal input of the processor device 12 or by analog-to-digital conversion of the complex-valued input signal in a preprocessing device which converts the analog-to-digital converted values of the processor device 12 to be obtained. For example, a predetermined number of samples may be a transmission duration of a data packet of a wireless communication standard multiplied by a sampling rate, or parts or multiples of the transmission duration of a data packet of a wireless communication standard multiplied by a sampling rate. Thus, for example, a data packet sent via Bluetooth can have a length of 625 bits. If a time window of, for example, 272 μs is sampled at a rate of 110 MHz, then, for example, 30,000 sample values can be obtained.

The processor device 12 includes a real part image 38 and an imaginary part picture 42 , The real part pictures 38 is designed to obtain real parts of the samples based on which in a real part path 44a Arithmetic operations can be performed. The imaginary part pictures 42 is adapted to obtain imaginary parts of the samples based on which in an imaginary part path 44b Arithmetic operations can be performed.

The processor device 12 is designed to be based on the obtained real parts in a potentizer 46a to form a normalization power of the real parts with an exponent of 2. The normalization power may be based on the exponent of the power of the component or chosen independently. For example, the normalization power may correspond to the exponent of the power divided by two. The processor device 12 is further configured to be based on the obtained real parts in a potentizer 48a To obtain the power with the exponent of greater than 2, so that a plurality of potentiated real parts is obtained.

The processor device 12 further comprises a summer 52a and a summer 54a in the real part path 44a , The summer 52a is designed to sum up the normalization powers and to obtain a summer of the normalization powers. The summer 52a is further configured to divide the obtained sum by the predetermined number of samples.

The summer 54a is designed to obtain a sum of the plurality of amplified samples, ie the real parts. This means that the power is based on the sum supplied by the summer 54a is delivered can be obtained. The summer 54a is further configured to divide the sum obtained by the predefined number of samples. The divided sum of the totalizer 54a may be an average of the samples, ie the real parts, if the received signal has an average value. If an interference-free signal, for example, is transmitted without a mean value, as can be done, for example, in the case of the wireless communication standard, then a mean value obtained on the receiver side can be interpreted to mean that the transmitted signal has been subject to interference. The summers 52a and 54a are further configured to determine a deviation of a raised real part or normalization quotient from the respectively obtained average value. In other words, a deviation based on the power can be obtained.

The real part path 44a further comprises a quotient device 56a , which is designed to be a real-part quotient based on the sums of the summers 52a and 54a to build. In other words, the sums of the summers 52a and 52b be set in relation to each other. For example. By dividing the power by the normalization power, a normalization of the power can be obtained. A normalization power with a value of 2 enables normalization of the power with respect to the power of the received signal, so that the result based on the real part quotient can be independent of the signal power. Alternatively, the real part quotient may, for example, also include a division of the normalization power by the power.

The real part path 44a includes a squarer 58a which is adapted to form a square value based on the real part quotient, ie to square the real part quotient.

The imaginary part path 44b is equivalent to the real part path 44a constructed and includes a potentizer 46b , which is designed to form the normalization powers of the imaginary parts, a potentizer 48b which is adapted to form the powers based on the imaginary parts, a summer 52b and a summer 54b , which are designed to add the powers or normalization powers or to form a mean value or to determine a deviation between the powers or normalization powers and the corresponding mean values. The imaginary part path 44b further comprises a quotient device 56b , which is designed to be an imaginary part quotient based on the sums of the summers 52b and 54b to build. The imaginary part path 44b includes a squarer 58b which is adapted to form a square value based on the imaginary part quotient, ie to square the imaginary part quotient.

The potentizers 46a and 48a , the summers 52a and 54a and the quotient facility 56a form a calculation path 65a , Furthermore, the potentizers form 46b and 48b , the summers 52b and 54b and the quotient facility 56b form a calculation path 65b , Alternative embodiments have alternative components of the calculation paths, as shown in subsequent embodiments.

The processor device 12 includes an adder 62 , which is adapted to the square of the first and the square of the second squarer 58a and 58b so that the intermediate values of the real part path and the imaginary part path are merged and a sum of the square values can be obtained. The processor device 12 further comprises a root image formed to be a root of the sum of the adder 62 where the root value is the result of the comparison device 14 is delivered. By forming the root can be a dimension of the quotients obtained from the quotient images 56a and 56b to be recovered by forming the square values in the squarers.

Alternative embodiments show devices that have all or part of the real part path or the imaginary part path. These embodiments have partially modified real subpaths or imaginary subpaths that produce the result from the respective quotient images 56a or 56b receive. By the absence of the other path 44a or 44b can be a squaring in the squarer 58a or 58b , as well as the summation in the adder 62 and rooting in the root image 64 omitted.

Compared to these embodiments has device 40 to the advantage that, by evaluating the real parts and imaginary parts of the samples, a sample of the sampled values which is larger than a view of the real parts or imaginary parts can be considered, ie evaluated. In return, a computing effort of the processor device 12 be reduced when the processor device 12 includes only the real part path or the imaginary part path wholly or partially.

Alternative embodiments show a normalization power that is within a tolerance of less than or equal to ± 10%, ± 15% or ± 20% with respect to the exponent of the power of the component divided by two. If the power of the component is, for example, 3, then the normalization power can lie at a tolerance of ± 10%, for example in a range between 1.35 and 1.65. For example, if the power of the component is 4, such as in a kurtosis calculation, the normalization potency may range between 1.6 and 2.4 at a tolerance of ± 20%.

In other words, device can 40 enable interference detection regardless of the overall power in the band since the signal power can be removed from consideration by the power normalization.

5 shows a schematic representation of a device 50 that unlike the device 40 only has a signal path. The device 50 has a bandpass filter 22 ' which is configured to filter a sample, ie a component or a sample, of the received signal and to obtain a filtered sample. The bandpass filter 22 ' can have a variable bandwidth, as is the case for the Bandpass filter 22 has been described. Alternatively, the bandpass filter 22 ' have a fixed bandwidth if, for example, a wireless communication in a fixed frequency spectrum with respect to interference is to be monitored. contraption 50 also has a potentizer 48 configured to provide a power of the filtered sample. contraption 50 has a summer 54 which is adapted to add up the powers of a plurality, ie the defined number, of samples and to deliver the result.

If only the real part or the imaginary part or the complex-valued received signal for the interference detection is evaluated, a summation of individual powers can be compared with the predefined threshold value. For example. For example, a summation value other than zero may indicate a DC component caused by interference on the transmission channel of the received signal.

6 shows a schematic representation of a device 60 for the detection of signal interference. The processor device 12 comprises a pre-processing device 66 , which is designed to pre-process a receiving signal. The preprocessing device 66 can, for example, the bandpass filter 22 , an optional preamplifier for amplifying the received signal and / or an analog-to-digital converter to digitize the received signal. Is device 60 For example, a Bluetooth device, bandpass filtering may, for example, filter a frequency range between 2402 and 2480 MHz.

Is device 60 configured to detect a Bluetooth interference for a Bluetooth environment, device may 60 also be referred to as mutual Bluetooth interference detection using higher order statistics.

contraption 60 further comprises an alarm device 74 configured to provide an alarm based on the signal interference result. For example, the processor device 12 or the comparison device 14 provide a non-zero signal value if the signal interference result indicates that the result of the processor device 12 is below the predefined threshold. If the comparator 14 outputs a signal interference result that indicates an exceeding of the threshold value and thus a signal interference, the comparison device can output a lower signal value of, for example, 0. The alarm device 74 For example, it can trigger an alarm if the signal value is reduced or set to 0. This variant has the advantage that a so-called zero-dead functionality is feasible, which can also generate an alarm, for example, when the processor device 12 or the comparator 14 has a defect and the processor device 12 no signal, ie a signal value of 0, outputs. Alternatively, the comparison device 14 For example, be configured to indicate the exceeding of the predefined threshold by a signal value that is different from 0 or greater than a default value, so that the alarm device 74 based on the 0 different or increased signal value can embarrass an alarm signal.

The processor device 12 is designed to calculate, based on the real part and the imaginary part of the preprocessed received signal, the fourth central moment, the so-called kurtosis, also called curvature of an empirical frequency distribution. contraption 60 includes the comparison device 14 , which is designed to compare the results with the predefined threshold. The processor device 12 includes the calculation paths 65a and 65b which are configured to calculate a power with an exponent of 4 and an exponent of 2 based on the real part and the imaginary part of the received signal, respectively, and form quotients based on the exponent of 4 and a square value of the exponent of 2.

In other words, the processor device 12 designed to calculate the kurtosis, ie the fourth central moment of the real part and the imaginary part. A value of kurtosis may be dependent on a magnitude of interference that occurs, so that, unlike traditional energy detectors that measure energy within a particular frequency band, any interference that may occur can be characterized by the use of kurtosis and deduced from the SINR.

wobei kurt(X) die Kurtosis, E[(X – μ)] den Erwartungswert einer Abweichung eines Wertes X von einem Mittelwert μ bezeichnet. Vorteilhaft an der Kurtosis ist, dass sie unabhängig von der Signalleistung ist und lediglich abhängig von der Kontur der PDF ist. Damit kann Vorrichtung 60 als SINR-basierter Interferenzdetektor bezeichnet werden.The kurtosis is mathematically defined as the standardized fourth moment of the mean and can be expressed by the following equation:

where kurt (X) denotes the kurtosis, E [(X-μ)] the expected value of a deviation of a value X from an average μ. The advantage of the kurtosis is that it is independent of the signal power and only depends on the contour of the PDF. This can be device 60 be referred to as SINR-based interference detector.

Furthermore, the kurtosis allows the cancellation of additive white gaussian noise (AWGN) effects. The kurtosis of a Gaussian random variable is equal to the value 3 according to formula 2. If the value of 3 is subtracted from the result, the kurtosis can assume the value 0 and thus the AWGN is neglected for further considerations.

It is understood that the illustrations and explanations are mathematical in nature. A device for detecting signal interferences can implement a realization of identical, equivalent or similar relationships. However, the formulas and relationships allow a representation of why the kurtosis is useful as a decision criterion for the presence of a signal interference and possibly advantageous.

The kurtosis β of a signal I is defined as

beschrieben werden kann.A mathematical empirical value of a value series is defined as the averaging over all values, ie at all times and therefore from -∞ to + ∞. contraption 60 is designed to estimate the respective empirical value over a finite number of N sampling points, so that the kurtosis of N signal values recorded as

can be described.

The estimator is unbiased in this case.

The estimated expectation E may also be used as a moving average, i. H. be carried out as an average of the received signal over a limited past time window, for example the length N. A consideration of newer, d. H. Present or future values may be realized by, for example, forgetting, i. H. it always takes into account a certain number of consecutive past values. Older values, for example, are disregarded. Alternatively, a weighting of past values can be carried out in such a way that a number N of past sampled values, possibly in the form of the past estimated empirical value, is weighted with a smaller factor than subsequent, new sampled values. Thus, for example, slow changes in the communication environment can be taken into account, as the past values can be further taken into account.

dargestellt werden. Eine Herleitung des Zusammenhangs ist am Ende der Beschreibung angefügt.Is the kurtosis, for example, an interfered signal of the form I (t) = A ₁ X ₁ (t) + A ₂ X ₂ (t) + N (t) (4) applied, the kurtosis according to formula 3 as

being represented. A derivation of the relationship is added at the end of the description.

In this case, X ₁ (t) and X ₂ (t) denote the useful signal or the interfering signal with transmission powers normalized to 1. N (t) denotes additive white Gaussian noise with a fixed variance σ 2/1 , Furthermore, it can be assumed for the following statements that the fourth order moments of the useful signal (m _1.4 ) and of the interferer (m _2.4 ) are fixed.

Propagation of radio signals over a non-line-off-sight (NLOS) can be considered via a Rayleigh fading channel over the Rayleigh distributed random variables A ₁ and A ₂ . These are represented in formula 5 in the form of p ₁ and p ₂ as parameters of the random variables A ₁ and A ₂ and allow to characterize differences in the transmission powers between the signals X ₁ and X ₂ .

The SINR with X ₁ as the useful signal and X ₂ as the interferer can be expressed as:

A comparison of formulas 5 and 6 shows that SINR is directly related to kurtosis. The SINR is based on the variables p ₁ , p ₂ and σ _n . These are also included in formula 6. This can be deduced by determining the kurtosis on the SINR and thus the presence or absence of interference. The predefined threshold value, for example a kurtosis, can be determined, for example, by a simulation or metrological detection, wherein a kurtosis value then permits a conclusion to a previously recorded SINR value.

This may also allow the intensity of interference to be determined by the presence of interference. For example, a SINR threshold for a decision to generate an alarm may depend on a required or desired bit error rate (BER). For example, for Bluetooth systems a minimum SINR of 11 dB may be sought to achieve a bit error probability ≤ 0.1% for co-channel interference, as described for example in [6].

In other words, the detection of signal interference involves two steps. In a first step, profile curves that relate the kurtosis to the SINR are empirically generated and a respective SINR threshold selected based on a required or targeted bit error rate and in the threshold memory of the device 32 deposited. A second step can be called a classification step. The kurtosis of a received signal is determined and compared with the stored threshold to make a decision regarding the interference occurring based on the comparison.

The processor device 12 further comprises a squarer 58'a and a squarer 58'b , which are trained to make a magnitude square of the kurtosis of the real part and an amount square of the kurtosis of the imaginary part. The adder 62 is designed to add the two squares of sums. The root images 64 is designed to draw a root from the obtained sum. The obtained sum becomes the comparator 14 provided.

In other words, device sets 60 on conventional monitoring and alerting systems. By considering the kurtosis existing limitations are overcome. Thus, the detection of interferences by signals and / or interferers can also be made possible below the typical transmission power of the primary systems, for example if the transmission power of the interferer and the useful signal are added up below an energy threshold value.

contraption 60 can also be referred to as an automatic monitoring and alerting system, which can detect interference, for example due to malfunction of individual radio systems, interference and external systems, and thus can open up the possibility to respond to the interference early on. As a result, the reliability and the time behavior of radio-based communication can be secured and / or its downtime can be reduced or minimized.

In other words, it is a device 60 designed to receive the electromagnetic waves with the antenna. Subsequent preprocessing (bandpass filtering, frequency conversion, etc.) ensures that only signals from systems installed in the environment are examined. Since so-called coexistence management is carried out in industrial plants, the information about all installed systems with bandwidth and center frequency is available and can be read from a database. In the subsequent processing step, the signals are split into in-phase and / or quadrature components. Subsequently, the calculation of the fourth central moment (kurtosis) takes place. The subsequent threshold decision is the interference detection. The detector's task is to increase safety and reliability. For this he compares the obtained data with stored values (threshold-based approach). In this case, the threshold value can be a value determined empirically for each system and stored in a database (high accuracy) or a fixed radio system-independent value. Another possibility is the use of a machine learning method, which receives as input the statistically obtained data.

Alternatively, it is also conceivable that the processor device 12 is designed to add the kurtosis of the real part and the kurtosis of the imaginary part without forming an absolute square and / or without the formation of a root of the addition of the comparator 14 provide. The formation of the squares and / or the calculation of the root can be described as a possibly reversible mathematical operation, possibly the comparison operations of the comparator 14 simplify. Will be the result of the processor device 12 Based solely on the evaluation of the real or imaginary parts of the complex received signal, a squaring and / or magnitude formation by the squarer 58'a or 58'b , an addition by the adder 62 or a root calculation by the root image 64 omitted, so that the comparison device 14 can be formed, the result of the quotient device 56a or 56b to compare with the predefined threshold. Since a relevant to predominant portion of the interference information can be obtained by calculating the kurtosis, it is conceivable that the threshold value is compared directly with the kurtosis values.

7 shows a schematic representation of a device 70 for the detection of signal interference. contraption 70 includes a processor device 12 ' , which is designed to form a real part of the preprocessed complex-valued received signal and to form the kurtosis of the real part. contraption 70 or the processor device 12 ' also has the comparison device 14 which is adapted to the kurtosis of the real part with a in the threshold memory of the device 32 stored threshold value to compare.

Compared to a calculation of the kurtosis for the real part and the imaginary part, as it is in 6 is explained, the formation of kurtosis only the real part to a reduced computational cost of the processor device 12 ' and thus lead to cheaper and / or more energy-efficient components. A calculation of the real part and the imaginary part may allow an increased calculation accuracy since the examined sample, ie a number of sample points, is increased or doubled. An enlargement of the sample can lead to a lower estimation error of the potency, such as the kurtosis.

8th shows a schematic representation of a simulative structure for obtaining a profile, ie of thresholds, which can be stored in threshold value stores of an apparatus for interference detection. A setup or setup as it is in 3 : shown can be used to obtain a SINR kurtosis profile.

The simulative structure may, for example, be implemented in a MATLAB environment and includes two signal sources 76a and 76b , which are designed to be based on a signal description, as shown in formula (4) and 14 is shown to generate the signals X ₁ and X ₂ . N (t) describes additive white Gaussian noise with the variance σ 2 / n

For the sake of simplification, it can be assumed that X ₁ and X _{2 are} independent random processes (mean-free) with normalized variances and fourth-order moments m _1.4 and m _2.4 , respectively. Furthermore, A ₁ and A _{2 are} Rayleigh distributed random variables with the parameters P ₁ and P ₂ .

The signals X ₁ and X ₂ can simulate Bluetooth or WLAN signals, for example. A noise generator 78 is designed to generate additive white Gaussian noise N. The structure is also designed to superimpose the signals X ₁ and X ₂ with Rayleigh coefficients A ₁ and A ₂ , thus simulating NLOS transmission links. The superposition of the signals X ₁ and X ₂ with the Rayleigh coefficients A ₁ and A ₂ allows the consideration of positions of Nutzsignalsendern or interferers and local conditions, such as a factory with different locations of machines, walls, doors and / or floors can be configured. In particular, existing NLOS connections between a transmitter and a receiver of the radio link can be modeled by the Rayleigh variables. The structure comprises two calculation paths 82a and 82b : A non-interfering white noise N computation path superimposed 82a with the useful signal X ₁ and one and an interfering computation path 82b with the interference signal X ₂ for the determination of the SINR.

In the interference-free calculation path 82a the useful signal X _{1 is influenced} by the Rayleigh variable A ₁ and superposed with the noise signal N, ie the noise signal N is added to the signal A ₁ × X ₁ . For this interference-free complex-valued signal, the real part and the imaginary part respectively Potencies, such as the kurtosis calculated and the results squared, added and pulled from the result a root, as for example in 6 for device 60 for interference detection is shown. Based on the information that the result obtained is interference-free, for different times and / or Rayleigh variables A _1, different setpoint values (k ₀ , k ₁ ,...) In a result block 83 be derived.

The interfered calculation path 82b includes a calculation block 84 which is adapted to receive, in addition to the Rayleigh-scattered signal A ₁ .X ₁ and the overlying white noise N, also the Rayleigh-scattered signal A ₂ .X ₂ representing an interferer such that the calculation block 84 the signal I (t), as described for example in formula 4, can be entered in other words, the calculation block 84 receive the signal I (t). The calculation block 84 is configured to calculate the SINR from the interfered signal I (t), as defined in Formula 6, for example. This can be a result block 86 , ie (SINR ₀ , SINR ₁ ,...) for the same scenarios as indicated by A ₁ and X ₁ in the interference-free signal path 82a are taken into account.

In other words, a direct comparison of an interference-free transmission for different times and / or scenarios in the result block 83 with the SINR values of the result block 86 compared to derive the thresholds.

Thus, a determination of an SINR value during a current operation of a device for interference detection omitted, since for such a device 10 a current scenario, ie the current parameters A ₁ , A ₂ , X ₁ , X ₂ and N are at least partially unknown. This means that a direct determination of the SINR during operation can be conditionally to unworkable, since transmission power and similar parameters are at least partially unknown. A calculation of the signal interference result based on the complex input signal, however, is feasible.

In other words, the MATLAB simulation overlays Bluetooth signals in an AWGN channel with Rayleigh scattering to model interference (for example, Bluetooth interference). The Bluetooth signals can be generated in a complex baseband. The kurtosis calculation takes into account the in-phase and quadrature components of each signal. The Bluetooth signal may have a data rate of 1 Mbps without spreading, resulting in a bit duration of approximately 1 μs. A data packet with a length of 625 bits is assembled for the simulation. For a simulation, it can additionally be assumed that the packets are transmitted in a single time interval (single-time slot packets). A time window of approximately 272 μs, which corresponds to approximately 30,000 sampling points, becomes in the respective detection process of the calculation path 82a and 82b to determine a SINR and kurtosis to allow for graphical output. For example, it is believed that over these 30,000 sample points, there is either full period interference, ie full overlap, or no interference, ie, no overlap. Alternative scenarios may also take into account a partial overlap, ie interference during part of the period of consideration.

A ₁ and A ₂ can be determined for experimental setup as random variables of Rayleigh scattering with A ₁ having a parameter of 0 dB and A ₂ with a parameter varying between 0 and 20 dB. The thermal noise power N is normalized to a value of 0 dB. For each SINR step, a certain number, for example 5,000, 10,000 or 20,000 simulations are executed. For further processing, for example, the median value or the geometric mean of the results of the total number of simulations can be taken into account.

9 shows a schematic representation of an experimental setup with a device 90 with laboratory equipment for the determination of predefined thresholds.

Bluetooth Bluetooth interference scenarios can be tested and sent to a software-defined radio environment (Software Defined Radio SDR). For this purpose, two Bluetooth sources without a direct line-of-sight (NLOS) are interfered on a same channel, whereby a transmission power of a transmitter is kept constant and a transmission power of the second signal generator is varied.

The device 90 is designed to, for example, a simulation environment, as in 8th is shown to mimic in hardware. The results may be used in addition to or as an alternative to the simulation, for example, to the results of the simulation and the device 90 compare or verify. The device 90 includes two signal generators 88a and 88b , The signal generators may, for example, be of the type SMIQ06B from Rohde & Schwarz and are designed to generate signals such as Bluetooth or WLAN signals with center frequencies of 2.420 GHz. SMIQ06B signal generators have one Frequency range from 300 kHz to 6.4 GHz and can generate signals with a transmission power of up to 16 dBm. A receiving device 92 is designed to be that of the signal generators 88a and 88b to receive over a modeled Rayleigh channel, so that at the receiving device 92 an interfered signal A ₁ X ₁ + A ₂ , X _{2 is} received, which is generated by the signal generators 88a and 88b and / or from the receiving device 92 of thermal noise N may be superimposed, so that at the receiving device 92 the modeled signal I (t) can be evaluated. The receiving device 92 For example, a USRP2 RFX2400 transceiver can be connected to a PC 94 connected is. On the PC 94 Software such as MATLAB and / or Simulink can be used to evaluate the received signals. The receiving device 92 is designed to shift signals to baseband and sample at a sampling rate of, for example, 100 mega samples per second (Msps). For a signal processing, for example in MATLAB, for example, 30,000 sampling points can be taken into account at a time. Furthermore, a suitable downsampling factor can be applied to adjust the sampling rate, for example to a simulation environment as described in US Pat 8th is shown.

10 shows a qualitative graph in which the abscissa indicates a SINR in dB and the ordinate a respective determined kurtosis value. The graph has two rows of data. A first data series, represented by triangular data points, shows possible results of a simulative kurtosis determination, as can be determined, for example, by a simulation as described in 8th is shown.

A value determination can be obtained, for example, so that the power of one source is kept constant for both curves and the power of the other source is varied. Upon receipt of the signal, kurtosis values are determined by in-phase and quadrature components and then combined for different SINR values in MATLAB. This will work for both the simulation, as in 8th is shown as an example, and the real-time setup, as shown in 9 is shown performed. 10 shows how kurtosis varies as a function of SINR for both cases.

A second data series represented by diamond-shaped data points shows possible results of an experiment with hardware components, as obtainable for example by an experiment, as described in US Pat 9 is shown. The graph shows that simulative and experimental results can essentially agree, and that as the SINR increases, that is, with a channel disturbed by decreasing interference, a value of potency, such as kurtosis, decreases. A result calculation based on a different potency from the kurtosis can lead to an increase in the SINR result value. A monotonic progression of a rising or falling curve over an increasing SINR can allow derivation of a respective existing threshold value.

In other words, the fourth central moment, the kurtosis, also changes depending on the signal to interference ratio. The kurtosis has significant differences in a range of SINR up to, for example, 10 dB; H. for example, it may vary between 3.2 and 2.6, which may allow detection of interference based on the kurtosis. The range of a SINR of 0 dB, i. H. Interference and useful signal have approximately the same power, up to a SINR of 11 dB can be classified as relevant. For example, with Bluetooth, with a SINR greater than 11 dB, a bit error probability of less than 0.1% can be achieved, which may be sufficient for many applications. Alternative embodiments show requirements which, for example, require a relevant range of up to 15, 17 or 20 dB and thus define such a relevant range. Also, for example, when viewing or monitoring another wireless transmission medium, such as WLAN or Wireless Metropolitan Area Network (WMAN), various values of a power value, such as a kurtosis value, may be obtained depending on a particular SINR.

After such a profile is obtained, a threshold point on the curve can be selected. This threshold can be used, for example, to classify received signals as being interfered or not interfering (based on the sensitivity of the receiver). For example, if a minimum SINR of 11 dB is needed to achieve a bit error rate of ≤ 0.1% for co-channel interference, as described in [6], for example, this may have a kurtosis of about 2.5 on the profile of the Correspond to graphs. In other words, each time a number of samples, for example with 30,000 received sample points with a kurtosis of more than 2.5, can not be considered as interfering and it can be assumed that a correct demodulation of the signal with a bit error probability ≤ 0.1% can be achieved. A received sample with a kurtosis of less than 2.5 may be considered erroneous if, for example, the bit error probability does not meet the requirements.

11 shows a schematic flow diagram of a method 100 for deciding whether a signal can be classified as being interfered, ie corrupted or not interfered.

In one step 102 For example, a received analog signal is digitized with an analog-to-digital conversion. In one step 104 the kurtosis of the real and / or imaginary part of the digitized signal is determined. In one step 106 a threshold value is provided, for example by requesting a corresponding kurtosis value from a memory module or a database for a required or desired SINR. In one step 108 will be in the step 104 certain potency, for example the kurtosis, with the step 106 compared to the threshold value. In one step 112 is based on the result of the comparison in step 108 a decision is made, for example, that a signal level indicating the signal interference result is reduced or increased. This may indicate to an alarm block to trigger an alarm if the decision indicates that the received signal is so heavily interfered that it is considered faulty.

12 shows a graph with five data series 96a e. The abscissa of the graph plots the interferer-to-noise ratio (INR) in dB. The ordinate indicates a probability of misdetection, that is, a misjudgment, for example, whether a signal interference is erroneously detected or not detected, for a respective INR. The different data series 96a -E differ in their respective SNR. data series 96a shows an SNR of -2 dB, data series 96b an SNR of 0 dB, data series 96c an SNR of 2 dB, data series 96d an SNR of 4 dB and data series 96e an SNR of 6 dB. In other words, a signal to signal to noise ratio increases from the data series 96a to the data series 96e , About the abscissa increases the INR for a data series, colloquially expressed, the disturber is "louder", ie the interferer sends with increasing transmission power. As the INR increases, the likelihood of a mis-decision increases as the interferer may increasingly be identified as the payload transmitter. At the same time, a probability of misdetection decreases with increasing SNR of the signal, ie with increasing signal quality the probability of a wrong decision can be reduced.

In other words shows 12 the probability of misjudgment for different SNR values in a performance test. For different SNR values, the INR is increased. A large number of iterations, for example 5,000, 10,000 or 20,000, can be executed per INR to provide a graph as exemplified in FIG 12 is shown to produce.

13 shows a graph with three data series 98a -C, which compares possible kurtosis values (plotted on the ordinate) for varying radio standards with varying values of abscissa-indicated SINR values. The data series 98a shows a disturbance, ie interference, of a Wifi signal with a Wifi signal. The data series 98b shows the kurtosis for a Wifi signal, which is disturbed by a Bluetooth signal. The data series 98c shows a course of kurtosis for a Wifi signal, which is disturbed by a Zigbee signal. Zigbee is a wireless standard that can be used, for example, for industrial sensor networks and can bridge wireless transmission distances of up to 100 m, for example.

Wifi signals are for example WLAN signals, which are on the 802.11 put on. For small to very small SINR values, ie values between -50 and -30 dB of the SINR, the kurtosis is approx. 1.8, which may correspond, for example, to a WLAN signal that is almost or completely uninterfered. As SINR increases, kurtosis increases for the data series 98a -C steadily to reach at about 0 dB a respective maximum, which may be a global maximum, and then fall off for further increasing SINR values. For data series 98a For example, if the SINR is sufficiently large, the kurtosis may drop from, for example, greater than 30 dB to almost the value for the respective transmitter and interferer transmission system, ie the WLAN system. For the data series 98b and 98c Thus, for Bluetooth or Zigbee interferers, the kurtosis also falls sharply, but in each case on the kurtosis value of the transmission system with the highest reception power. This can be influenced by power losses that occur on different paths, ie transmission paths, between transmitter and receiver. This can lead to a reception power of a ZigBee signal being greater than a Bluetooth signal despite a lower transmission power. This means that Zigbee can have a higher reception performance than WLAN and Bluetooth a higher reception performance than Zigbee. For example. A WLAN signal with a transmission power of up to 60 mW, a Zigbee signal with a transmission power of up to 10 mW and Bluetooth with a transmission power of up to 100 mW can be transmitted. A gray backed area 102 of the graph describes an area covered by a device, such as device 10 . 20 . 60 or 70 a process such as the process 100 , so that can be classified as a result that has a value within the range 102 has been classified as overly interfered. This can be a threshold 104 be predefined. In other words then the signal can be considered erroneous, since the kurtosis values are above the possible threshold 104 lie. A kurtosis of about 2.2 indicates a possible threshold 104 ,

An interference detection based on the kurtosis can be particularly good at about the same size of useful signal transmitter and interferer, d. H. at a SINR of about 0 dB, since there are large variations in the kurtosis. This area can be described as a critical or the most critical area in terms of telecommunications, since demodulation of the useful signal is often or usually prevented here. Detection of interference in this area can therefore offer great advantages, especially if signal interferences are not or poorly detected by means of performance consideration of useful and interference signals of energy detectors. This can happen, for example, if the transmission power of the useful signal transmitter and the interferer are low and a common transmission power remains below a threshold value. In contrast, the performance independence of the kurtosis allows interference detection even in such cases.

14 shows a schematic representation of a system model for the description of the received signal I (t). A radio signal X ₁ (t) is multiplied by the Rayleigh variable A ₁ to model the transmission of the signal X ₁ (t) over a Rayleigh channel. A second signal X ₂ (t) is accordingly multiplied by a Rayleigh variable A ₂ to model the propagation of the second signal X ₂ (t) over a second channel. The two NLOS signals are in one point 106 to represent the joint transmission of signals X ₁ (t) and X ₂ (t) over a common channel. In one point 108 the sum of thermal noise N (t) is added up to obtain the signal I (t), as described for example in formula 4.

15 shows a schematic representation of channels, as occupied for example by Bluetooth in the 2.4 GHz ISM band. In a frequency range between 2.402 and 2.480 GHz are at a distance 112 of 1 MHz, a number of 79 channels, each with a bandwidth 112 1 MHz, with gaps between the channels for reducing or reducing channel cross talk. In other words, Bluetooth's 2.4GHz ISM band is divided into 79 subchannels. The entire or partially identical ISM frequency band is used by other radio standards, for example 802.11 b / g (WLAN) with a different channel distribution. Simultaneous use of the medium through different, possibly mutually uncoordinated, standards can lead to the occurrence of interference.

Derivation of the kurtosis description

The kurtosis β of a signal I can be defined as:

The fourth-order moment of I, which is also mean-free, since it is a sum of mean-free random variables, is defined by m _{I, 4} = E [I ⁴ ]. For I ⁴ applies: I ⁴ = (A ₁ X ₁ + A ₂ X ₂ + N) ⁴ = (A ₁ X ₁ + A ₂ X ₂ + N) ² (A ₁ X ₁ + A ₂ X ₂ + N) ² = A 4 / 1X 4/1 + A 2 / 1A 2 / 2X 2 / 1X 2/2 + A 2 / 1X 2 / 1N ² + 2A 3 / 1A ₂ X 3 / 1X ₂ + 2A 3 / 1X 3 / 1N + 2A 2 / 1A ₂ X 2 / 1X ₂ N + A 2 / 1A 2 / 2X 2 / 1X 2/2 + A 4 / 2X 4/2 + A 2 / 2X 2 / 2N ² + 2A 3 / 2A ₁ X 3 / 2X ₁ + 2A 3 / 2x 3 / 2N + 2A 2 / 2A ₁ X 2 / 2X _1N 2A 3 / 1A ₂ X 3 / 1X ₂ + A 2 / 1X 2 / 1N ² + + A 2 / 2X 2 / 2N ² + N ⁴ + 2A ₁ A ₂ X ₁ X ₂ N ² + 2A ₁ X ₁ N ³ + 2A ₂ X ₂ N ³ + 2A 3 / 2A ₁ X 3 / 2X ₁ + 2A ₁ A ₂ X ₁ X ₂ N ² + 4A 2 / 1A 2 / 2X 2 / 1X 2/2 + 4A 2 / 1A ₂ X 2 / 1X ₂ N + 4A 2 / 2A ₁ X 2 / 2X ₁ N + 2A 3 / 1X 3 / 1N + 2A ₁ A 2 / 2X ₁ X 2 / 2N + 2 ₁ X ₁ N ³ + 4A 2 / 1A ₂ X 2 / 1X ₂ N + 4A 2 / 1X 2 / 1N ² + 4 ₁ A ₂ X ₁ X ₂ N ² + 2A 3 / 2X 3 / 2N + 2A ₂ A 2 / 1X ₂ X 2 / 1N + 2A ₂ X ₂ N ³ + 4A 2 / 2A ₁ X 2 / 2X ₁ N + 4A 2 / 2X 2 / 2N ² + 4A ₂ A ₁ X ₂ X ₁ N ² (7) if the terms of the same dimension are summed up, then the above equation results: I ⁴ = A 4 / 1X 4/1 + A 4 / 2X 4/2 + N ⁴ + 6A 2 / 1A 2 / 2X 2 / 1X 2/2 + 6A 2 / 1X 2 / 1N ² + 4A 3 / 1A ₂ X 3 / 1X ₂ + 2A ₁ A ₂ X ₁ X ₂ N ² + 4A 3 / 1X 3 / 1N + 6A 2 / 2X 2 / 2N ² + 4A ₁ A 3 / 2X ₁ X 3/2 + 4A 3 / 2X 3 / 2N + 4A ₁ X ₁ N ³ + 4A ₂ X ₂ N ³ + 10A ₁ A ₂ X ₁ X ₂ N ² + 12A 2 / 1A ₂ X 2 / 1X ₂ N + 12A 2 / 2A ₁ X 2 / 2X N ₁ (8)

Due to the linearity of the expected value operator, the expected value of I ^{4 is} the sum of the expected values of the individual components: E [I ⁴ ] = E [A 4 / 1X 4/1] + E [A 4 / 2X 4/2] + E [N ⁴ ] + 6E [A 2 / 1A 2 / 2X 2 / 1X 2/2] + 6E [A 2 / 1X 2 / 1N ² ] + 4E [A 3 / 1A ₂ X 3 / 1X ₂ ] + 2E [A ₁ A ₂ X ₁ X ₂ N ² ] + 4E [A 3 / 1X 3 / 1N ] + 6E [A 2 / 2X 2 / 2N ² ] + 4E [A 3 / 1A 3 / 2X ₁ X 3/2] + 4E [A 3 / 2X 3 / 2N] + 4E [A ₁ X ₁ N ³ ] + 4E [A ₂ X ₂ N ³ ] + + 10E [A ₁ A ₂ X ₁ X ₂ N ² ] + 12E [A 2 / 1A ₂ X 2 / 1X ₂ N] + 12E [A 2 / 2A ₁ X 2 / 2X ₁ N] (9)

X ₁ , X ₂ , A ₁ , A ₂ and N are independent of each other such that f ₁ (X ₁ ), f ₂ (X ₂ ), f ₃ (A ₁ ), f ₄ (A ₂ ), and f ₅ (N) for any functions f ₁ , f ₂ , f ₃ , f ₄ and f _{5 are} also independent of each other. So formula 9 could also be written as: E [I ⁴ ] = E [A 4/1] E [X 4/1] + E [A 4/2] E [X 4/2] + E [N ⁴ ] + 6E [A 2/1] E [A 2/2] E [X 2/1] E [X 2/2] + 6E [A 2/1] E [X 2/1] E [N ² ] + 4E [A 3/1] E [ A ₂ ] E [X 3/1] E [X ₂ ] + 2E [A ₁ ] E [A ₂ ] E [X ₁ ] E [X ₂ ] E [N] + 4E [A 3/1] E [ X 3/1] E [N] + 6E [A 2/2] E [X 2/2] E [N ² ] + 4E [A ₁ ] E [A 3/2] E [X ₁ ] E [X 3/2] + 4E [A 3/2] E [X 3/2] E [N] + 4E [A ₁ ] E [X ₁ ] E [N ³ ] + 4E [A ₂ ] E [X ₂ ] E [N ³ ] + 10E [A ₁ ] E [A ₂ ] E [X ₁ ] E [X ₂ ] E [N ² ] + 12E [A 2/1] E [A ₂ ] E [X 2/1 ] E [X ₂ ] E [N] + 12E [A 2/2] E [A ₁ ] E [X 2/2] E [X ₁ ] E [N] (10)

Since N, X ₁ and X _{2 are} mean free, Equation 10 reduces to: E [I ⁴ ] = E [A 4/1] E [X 4/1] + E [A 4/2] E [X 4/2] + E [N ⁴ ] + 6E [A 2/1] E [A 2/2] E [X 2/1] E [X 2/2] + 6E [A 2/1] E [X 2/1] E [N ² ] + 6E [A 2/2] E [ X 2/2] E [N ² ] (11)

A ₁ and A ₂ are Rayleigh distributed and have moment second and fourth order 2p 2/1 , 2p 2/2 and 8p 4/1 , or 8p 4/2 on. N has a normal distribution with second and fourth order moments of σ 2 / N , or 3σ 4 / N on. X ₁ and X ₂ have second order moments normalized to 1 and fourth order moments m _1.4 and m _2.4 , respectively. Therefore, E [I ⁴ ] reduces to: E [I ^4] = 8m _1.4 p 4/1 + 8m 4 p _2.4 / 2 + 3σ 4 / n + 2 24p / 1p 2/2 + 2 12σ / np 2/1 + 2 12σ / np 2 / 2 (12)

The second-order moment of I results in: E [I ² ] = E [(A ₁ X ₁ + A ₂ X ₂ + N) ² ] = E [A 2/1 X 2/1 + A 2 / 2X 2/2 + N ² + 2A ₁ A ₂ X ₁ X ₂ + 2A ₁ X ₁ N + 2A ₂ X ₂ N] (13) = E [A 2/1] E [X 2/1] + E [A 2/2] E [X 2/2] + E [N ² ] + 2E [A ₁ ] E [A ₂ ] E [X ₁ ] E [X ₂ ] + 2E [A ₁ ] E [X ₁ ] E [N] + 2E [A ₂ ] E [X ₂ ] E [N] (14)

Since X ₁ , X _2, and N are independent of each other and mean-free, Equation 14 reduces to: E [I ² ] = E [A 2/1] E [X 2/1] + E [A 2/2] E [X 2/2] + E [N ² ] = 2p 2/1 + 2p 2 / 2 + σ 2 / n (15)

A squaring of the term yields: (E [I ² ]) ² = 4p 4/1 + 4p 4/2 + 8p 2 / 1p 2/2 + 4σ 2 / np 2/1 + 4σ 2 / np 2/2 + σ 4 / n (16)

womit Gleichung 5 hergeleitet ist.Finally, to obtain the kurtosis, as shown for Equation A, the fourth-order moment is divided by the square of the second-order moment:

with which equation 5 is derived.

Although some aspects have been described in the context of a device, it will be understood that these aspects also constitute a description of the corresponding method, so that a block or a component of a device is also to be understood as a corresponding method step or as a feature of a method step. Similarly, aspects described in connection with or as a method step also represent a description of a corresponding block or detail or feature of a corresponding device.

Depending on particular implementation requirements, embodiments of the invention may be implemented in hardware or in software. The implementation may be performed using a digital storage medium, such as a floppy disk, a DVD, a Blu-ray Disc, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, a hard disk or other magnetic or optical memory are stored on the electronically readable control signals, which can cooperate with a programmable computer system or cooperate such that the respective method performed becomes. Therefore, the digital storage medium can be computer readable. Thus, some embodiments according to the invention include a data carrier having electronically readable control signals capable of interacting with a programmable computer system such that one of the methods described herein is performed.

In general, embodiments of the present invention may be implemented as a computer program product having a program code, wherein the program code is operable to perform one of the methods when the computer program product runs on a computer. The program code can also be stored, for example, on a machine-readable carrier.

Other embodiments include the computer program for performing any of the methods described herein, wherein the computer program is stored on a machine-readable medium.

In other words, an embodiment of the method according to the invention is thus a computer program which has a program code for performing one of the methods described herein when the computer program runs on a computer. A further embodiment of the inventive method is thus a data carrier (or a digital storage medium or a computer-readable medium) on which the computer program is recorded for carrying out one of the methods described herein.

A further embodiment of the method according to the invention is thus a data stream or a sequence of signals, which represent the computer program for performing one of the methods described herein. The data stream or the sequence of signals may be configured, for example, to be transferred via a data communication connection, for example via the Internet.

Another embodiment includes a processing device, such as a computer or a programmable logic device, that is configured or adapted to perform one of the methods described herein.

Another embodiment includes a computer on which the computer program is installed to perform one of the methods described herein.

In some embodiments, a programmable logic device (eg, a field programmable gate array, an FPGA) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor to perform one of the methods described herein. In general, in some embodiments, the methods are performed by any hardware device. This may be a universal hardware such as a computer processor (CPU) or hardware specific to the process, such as an ASIC.

The embodiments described above are merely illustrative of the principles of the present invention. It will be understood that modifications and variations of the arrangements and details described herein will be apparent to others of ordinary skill in the art. Therefore, it is intended that the invention be limited only by the scope of the appended claims and not by the specific details presented in the description and explanation of the embodiments herein.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 020050061057 A1 [0018]
• US 000008295189 B2 [0018]
• US 000008290100 B2 [0018]
• US 000007149205 B2 [0018]
• WO 002008112928 A2 [0018]
• US 020080192644 A1 [0018]
• US 020080279093 A1 [0018]
• US 020080299932 A1 [0018]
• JP 000S58133060A [0018]
• JP 000S581373337 A [0018]
• JP 000H08316996A [0018]
• JP 002002335196 A [0018]
• JP 002002353878A [0018]
• JP 002008054303 A [0018]
• JP 002009267855 A [0018]
• GB 000002451682 A [0018]
• CN 000101047486 B [0018]
• CN 000101692631 B [0018]
• CN 000101697624 B [0018]
• US 20090197556 A1 [0018]
• US 6130907 A [0018]
• US 6704378 B2 [0018]
• EP 2391022 A1 [0018]

Cited non-patent literature

• VDI / VDE Directive 2185 [0004]
• IEEE 802.11G [0052]
• Standards 802.11b, g or n [0062]
• 802.11 standard [0135]

Claims (16)

Device for detecting signal interference in a received signal, comprising: a processor device ( 12 ; 12 ' ), which is designed to be based on a power ( 48a ; 48b ) of a component ( 38 ; 42 ) of the received signal with an exponent greater than 2 to obtain a result which is based on the power ( 48a ; 48b ) is based; and a comparator ( 14 ) for comparing the result with a predefined threshold value ( 16 ) to obtain a signal interference detection result depending on the comparison. Device according to Claim 1, in which the processor device ( 12 ; 12 ' ) is further configured to be based on a normalization power ( 46a ; 46b ) of the component ( 38 ; 42 ) of the received signal with an exponent corresponding within a tolerance of ± 20% to a value of the power divided by two, to obtain a normalization result, and the result based on a quotient ( 56a ; 56b ) based on the power ( 48a ; 48b ) with the exponent greater than 2 and the normalization power ( 46a ; 46b ) to obtain. Device according to Claim 1 or 2, in which the received signal is a complex received signal, in which the processor device ( 12 ; 12 ' ) is designed to be a real part ( 38 ) or an imaginary part ( 42 ) of the complex received signal, and in which the processor device ( 12 ; 12 ' ) is further configured to calculate the result based on the power ( 48a ) of the extracted real part ( 38 ) or based on the power ( 48b ) of the extracted imaginary part ( 42 ) to obtain. Device according to Claim 3, in which the processor device ( 12 ; 12 ' ) is designed to calculate an amount ( 58b ' ) of potency ( 48b ) of the imaginary part ( 42 ) and an amount ( 58a ' ) of potency ( 48a ) of the real part ( 38 ) ( 62 ) to get a sum and calculate a root from the sum ( 64 ). Device according to one of the preceding claims, in which the processor device ( 12 ; 12 ' ) is designed to calculate the result based on the power ( 48a ; 48b ) with an exponent greater than or equal to 2.8. Device according to one of the preceding claims, in which the processor device ( 12 ; 12 ' ) is designed to generate a predetermined number (N) of samples of the component ( 38 ; 42 ) capture; to exponentiate each sample with the exponent ( 48a ; 48b ) to obtain a plurality of exponentiated samples and to sum the plurality of exponentiated samples using the number of samples ( 54a ; 54b ); to get a sum; and to get the power based on the sum. Device according to one of the preceding claims, in which the received signal is an averaged signal and in which the processor device ( 12 ; 12 ' ) is adapted to generate a predetermined number (N) of samples based on the component ( 38 ; 42 ) capture; to form an average of the samples ( 54a ; 54b ); to obtain a deviation between at least one of the samples and the mean; and to get the power based on the deviation. Device according to one of the preceding claims, further comprising a memory device ( 16 ), which is designed to store the predefined threshold value and the comparison device ( 14 ). Device according to Claim 8, in which the memory device ( 16 ) is further configured to store at least one further predefined threshold value and the comparison device ( 14 ) based on the received signal. Apparatus according to any one of the preceding claims, adapted to receive received signals based on different wireless communication standards, which also has a configurator ( 26 ) configured to be automatically or based on a user input ( 28 ) is configurable either for a first wireless communication standard or for a different second wireless communication standard, the processor means ( 12 ; 12 ' ) a bandpass filter ( 22 ), which has an adjustable bandwidth, wherein the configurator ( 26 ) is configured, in a configuration to the first wireless communication standard, the bandpass filter ( 22 ) to a first bandwidth and a first predefined threshold value of the comparison device ( 14 ) and, when configured to the second wireless communication standard, the band pass filter ( 22 ) to set a second, different bandwidth and a second, different predefined threshold value of the comparison device ( 14 ) to deliver. Device according to one of the preceding claims, further comprising an alarm device ( 74 ) configured to provide an alarm signal based on the signal interference detection result. Device according to one of the preceding claims, wherein the threshold value corresponds to a signal to interference ratio (SINR) of less than or equal to 11 dB, when the received signal is a Bluetooth signal. Device according to one of the preceding claims, in which the received signal is a complex signal and the processor device ( 12 ; 12 ' ) is adapted to detect a number (N) of samples of the complex received signal; To extract real parts and imaginary parts of the samples of the complex received signal ( 38 ; 42 ); the potencies ( 48 ; 48a ; 48b ) of the extracted real parts ( 38 ) and the extracted imaginary parts ( 42 ) with an exponent of 4; to obtain an average of the powers ( 52a ; 52b ); the normalization powers ( 46 ; 46a ; 46b ) of the extracted real parts ( 38 ) and the extracted imaginary parts ( 42 ) with an exponent of 2; an average of the normalization powers ( 54 ; 54a ; 54b ) to obtain; a real part quotient ( 56a ) based on the mean ( 54a ) of the potencies ( 48a ) of the real parts ( 38 ) and the mean ( 52a ) of the normalization powers ( 46a ) of the real parts ( 38 ), and an imaginary part quotient ( 56b ) based on the mean ( 54b ) of the potencies ( 48b ) of the imaginary parts ( 42 ) and the mean ( 52b ) of the normalization powers ( 46b ) of the imaginary parts ( 42 ) to obtain; a sum value ( 62 ) based on a square value ( 58a ), which is based on the real part quotient ( 56a ) and a square value ( 58b ), which is based on the imaginary part quotient ( 56b ); and the result based on a root value ( 64 ), which is based on the sum value ( 62 ). Method for determining a threshold value comprising the following steps: receiving a reference signal (A ₁ X ₁ ); Receiving an interfered signal (I (t)) based on the reference signal (A ₁ X ₁ ) and an interfering signal (A ₂ X ₂ ). Receive ( 82b ) the signal-to-interference-plus-noise-ratio (SINR) based on the interfered signal (I (t)); Receive ( 82a ) a result based on a power of a component of the reference signal having an exponent greater than 2; Comparing the result with the SINR; and storing the result in a threshold memory ( 16 ). Method for detecting signal interference in a received signal, comprising the following steps: Obtaining a result based on a power of a component of the received signal having an exponent greater than 2; and Comparing the result with a predefined threshold value to obtain a signal interference detection result depending on the comparison. A computer program comprising program code for performing the method of claim 15 when the program is run on a computer.

Images