WO2012104773A1

WO2012104773A1 - Lighting control system

Info

Publication number: WO2012104773A1
Application number: PCT/IB2012/050423
Authority: WO
Inventors: Dagnachew Birru; Shenqiu ZHANG; Yao-Jung WEN; Jianfeng Wang
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2011-02-04
Filing date: 2012-01-30
Publication date: 2012-08-09

Abstract

A method has been proposed to operate a lighting control system (100) consisting of photo sensors (110), dimmable electric light sources (120), a motorized shading system(130) and a controller (140). Initial illuminance data based on different environmental conditions such as exterior illuminance, blind height, and blind angle are received and classified. Estimators are calculated for these different environmental conditions. The illuminance in the taskplane is estimated. Based upon the estimated taskplane illuminance, the dimmable electric light sources (120) and the motorized shading system (130) are adjusted to meet a required setpoint.

Description

Lighting control system

FIELD OF THE INVENTION

The present invention relates to the control of lighting systems. Particularly it relates to methods and a system for controlling light based on illuminance estimation.

BACKGROUND OF THE INVENTION

Photo controlled lighting systems are nowadays being installed in all types of buildings for the purpose of reducing energy consumption and for saving costs. Such photo controlled lighting systems are typically employing light sensors or photo sensors. Lighting control systems are in general working according to a closed-loop control strategy by measuring the interior illuminance and adjusting the electric lights to meet a certain setpoint. The photo sensor for such systems is typically installed on the ceiling and is therefore measuring the ceiling illuminance. However, the main interest is to meet the desired set point at the workplane. Lighting control systems that are installed in a typical work space, such as an office area, are therefore preferably working to control the task lighting or task

illuminance on a work surface in the workplane.

Algorithms and configurations deployed today are mostly not robust enough to handle the impact of daylight conditions and window blinds or shades. In addition, daylight falling on the ceiling, when reflected from the blinds, can mislead the photosensor. In order to mitigate this, photo sensors are typically installed far away from the window and are covered or even placed in a hole so that the sensing element does not see the daylight directly. This can however result in more problems. Placing in a hole results in the sensor being sensitive to changes under its narrower field of view. Placing the sensor far away from the window makes the sensor unable to satisfactorily sense the normal daylight level in the room. This and other issues contribute to the shortcomings of today's solutions.

Several methods for daylighting applications have been developed. A method with an interior light sensor for the workplane illuminance estimation (prediction) in daylighting control systems is introduced in "Workplane illuminance prediction method for daylighting control systems " (Kwang-Wook Park, Andreas K. Athienitis, February 24, 2003). Park discloses a method for integrated control of motorized daylighting devices and dimmable electric lights. The workplane illuminance is predicted based on readings from one light sensor and different blind tilt angles of a motorized blinds system. Results show that the electric light workplane illuminance can be predicted with high accuracy using the method discussed in Park.

However, there is still a need for improved lighting control systems using advanced workplane illuminance estimation methods that are taking into consideration the impact of direct daylight, different light intensities, various sky conditions, window blinds or shades, changing interior and weather conditions.

SUMMARY OF THE INVENTION

In view of the above it is desirable that a lighting control system is operated in a more user acceptable and robust way. It is also desirable that a lighting control system is capable to predict the amount of light on the workplane based on different environmental conditions. This requirement implies that all types of conditions such as interior light, exterior light, blind height, and blind angle should be taken into account. In other words, it is desirable to consider a complex system setup of multiple light sources and multiple photo sensors and to use an advanced estimation based on environmental conditions to estimate the workplane illuminance.

In addition, it is desirable to be able to consider non diffused exterior light and different light intensities. Further it is desirable to consider many different possible combinations of blind tilt angles with blinds pulled down all the way or blinds pulled down to a certain height.

It is an object of the present invention to overcome these problems, and to provide methods and systems which solve or at least mitigate the issues addressed above.

Generally, the above objectives are achieved by methods and devices according to the attached independent claims.

According to a first aspect of the invention, the above objects are achieved by a method for operating a lighting control system comprising at least one photo sensor, at least one dimmable electric light source, a motorized shading system and a controller, the method comprising receiving initial illuminance data based on at least two environmental conditions n, wherein the at least two environmental conditions n are chosen from the group of exterior illuminance Ew, blind height h, and blind angle ω; classifying the initial illuminance data based on predetermined thresholds of the at least two environmental conditions; calculating estimators for the at least one environmental condition n; receiving measured illuminance data from at least one photo sensor; estimating a taskplane illuminance I_task for a current set of environmental conditions n by determining the taskplane illuminance I_task as a function of the calculated estimators and the measured illuminance data; receiving a taskplane illuminance setpoint; providing control signals to the at least one dimmable electric light source and the motorized shading system, thereby minimizing the difference between the taskplane illuminance setpoint and the estimated taskplane illuminance I_tas

The method according to the first aspect of the invention enables control of a daylight and electric light control system. The receiving and classifying of initial illuminance data based on different environmental conditions allow a large amount of data covering many different light conditions to be gathered and classified. This data is used to compute estimation relations between the amount of light at the taskplane and the signals of the photo sensors. Based on these estimation relations, the electric lights and the motorized shading system are controlled in order to achieve the desired taskplane illuminance.

The classifying of initial illuminance data may further comprise classifying data into K clusters by using the predetermined thresholds of the exterior illuminance E_lt E₂,^■■■ , E_K_₁, wherein the minimum and maximum possible value of the exterior illuminance is denoted as E₀ and E_K respectively.

The classifying of initial illuminance data may further comprise classifying data into L clusters by using the thresholds of blind height h , h₂, ··· , h_L_ , wherein the minimum and maximum possible value of blind height is denoted as h₀ and h_L respectively.

The classifying of initial illuminance may further comprise classifying data into M clusters by using the thresholds of blind angle ω₁, ω₂, ··· , ω_Μ_ι, wherein the minimum and maximum possible value of blind angle is denoted as ω₀ and ω_Μ respectively.

The data classification based on different combinations of exterior

illuminance, blind height and blind angle enables a large amount of data clusters to be gathered. There are many different possible clustering orders and levels, which permits extensive data classification, which in its turn generates a more precise estimation.

In an embodiment, the controller may comprise a light controller and a blinds controller and the providing control signals step may comprise providing control signals to the at least one dimmable electric light source from the light controller and providing control signals to the motorized shading system from the blinds controller. In contrast to an integrated controller, this may be an alternative where the control of the blinds and the electric lights are separated or independent. In an embodiment, the at least one photo sensor may be at least one taskplane photo sensor located at at least one point of interest in a taskplane and the lighting control system may further comprise a lighting actuator for the at least one dimmable electric light source, the method may further comprise the steps of turning off the at least one dimmable electric light source; receiving initial illuminance data from the at least one taskplane photo sensor; actuating the at least one dimmable electric light source to at least one prescribed lighting level; receiving illuminance data from the at least one taskplane photo sensor;

deriving, for the at least one point of interest, an illuminance model for the at least one dimmable electric light source being actuated to the at least one prescribed lighting level; calculating an electric light illuminance contribution, thereby separating the electric light illuminance contribution from an exterior illuminance contribution, for the at least one point of interest; receiving a desired illuminance setpoint for the at least one point of interest;

providing control signals to the at least one dimmable electric light source and the motorized shading system, thereby generating a desired task illuminance at the at least one point of interest.

These additional steps allow for forming of an illuminance model, where the illuminance measurement is associated with the corresponding light source and its output level. This allows the control system to gain knowledge of the relationship between electric light contribution and daylight contribution on a particular point of interest. This knowledge of net daylight on a particular point of interest may be used for the control of a shading system.

According to a second aspect, the above objects are achieved by a system comprising at least one photo sensor, at least one dimmable electric light source, a motorized shading system, a controller, wherein the controller is adapted to receive initial illuminance data based on at least two environmental conditions n, wherein the at least two environmental conditions n are chosen from the group of exterior illuminance Ew, blind height h, and blind angle ω; classify the initial illuminance data based on predetermined thresholds of the at least two environmental conditions n; calculate estimators for the at least two environmental conditions n; receive measured illuminance data from the at least one photo sensor; estimate a taskplane illuminance I_task for a current set of environmental conditions n by determining the taskplane illuminance I_task as a function of the calculated estimators and the measured illuminance data; receive a taskplane illuminance setpoint; provide control signals to the at least one dimmable electric light source and the motorized shading system, thereby minimizing the difference between the taskplane illuminance setpoint and the estimated taskplane illuminance.

In an embodiment, the controller may comprising a light controller and a blinds controller and the controller may further be adapted to provide control signals to the at least one dimmable electric light source from the light controller and to provide control signals to the motorized shading system from the blinds controller.

In an embodiment, the at least one photo sensor may be at least one taskplane photo sensor located at at least one point of interest in a taskplane, wherein the system is further comprising a lighting actuator for the at least one dimmable electric light source, wherein the controller is further adapted to turn off the at least one dimmable electric light source; receive initial illuminance data from the at least one taskplane photo sensor; wherein the actuator is adapted to actuate the at least one dimmable electric light source to at least one prescribed lighting level; wherein the controller is further adapted to receive measured illuminance data from the at least one taskplane photo sensor; derive, for the at least one point of interest, an illuminance model for the at least one dimmable electric light source being actuated to the at least one prescribed lighting level; calculate an electric light illuminance contribution, thereby separating the electric light illuminance contribution from an exterior illuminance contribution, for the at least one point of interest; receive a desired illuminance setpoint for the at least one point of interest; provide control signals to the at least one dimmable electric light source and the motorized shading system, thereby generating a desired task illuminance at the at least one point of interest.

It is noted that the invention relates to all possible combinations of features recited in the claims. Likewise, the advantages of the first aspect apply to the second aspect and vice versa.

BRIEF DESCRIPTION OF THE DRAWINGS

This and other aspects of the present invention will now be described in more detail, with reference to the appended drawings showing embodiments of the invention.

Fig. 1 illustrates a lighting control system according to an embodiment of the present invention.

Fig. 2 illustrates a system overview of a lighting control system according to an embodiment of the present invention.

Fig. 3 illustrates a system overview of a lighting control system according to an embodiment of the present invention. Figs. 4-6 illustrate flowcharts of data clustering algorithms according to embodiments of the present invention.

Fig. 7 is a schematic illustration of a method according to an embodiment of the present invention.

Fig. 8 is a schematic illustration of a method according to an embodiment of the present invention.

Figs. 9a-9f illustrates measured and estimated data according to embodiments of the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

The below embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout.

Fig. 1 illustrates a lighting control system 100 that may comprise at least one photo sensor 110, at least one electric light source 120, a shading system 130 and a controller 140. The lighting control system 100 may optionally have a lighting actuator 150. The at least one photo sensor 110 may be an interior photo sensor that may be wireless and installed on the ceiling, thus measuring the ceiling illuminance. However, as the most interesting is to measure the illuminance in the workplane, the illuminance sensed by this photosensor will be used to estimate workplane illuminance. An additional at least one photo sensor may be installed at a point of interest in the workplane or taskplane during initial calibration measurements. The point of interest in the workplane may be the surface of a desktop, a working table or similar. The at least one photo sensor 110 may be placed on other locations such as on the interior wall or on the places. The lighting control system 100 may optionally be combined with open loop blind control strategies for geography based glare control, whereby the system may optionally comprise a glare control photosensor. The at least one electric light source 120 may be a dimmable lighting fixture or luminaire. The shading system 130 may be a motorized blinds, roller shade system or other window treatment systems. The controller 140 may be an integrated controller that is providing control signals to both the at least one dimmable electric light source 120 and the motorized shading system 130. The controller 140 may optionally comprise two separate controllers, a light controller and a blinds controller, that are each controlling the at least one dimmable electric light source 120 and the motorized shading system 130 respectively. In the case of two separate controllers, the blinds controller may need to communicate its status to the workplane illuminance estimation system.

The lighting control system may optionally comprise a lighting actuator 150. The lighting actuator may be dimmable and addressable and may actuate an electric light source 120 to a prescribed dimming level.

The lighting control system may further comprise other devices such as a taskplane illuminance sensor being used for measuring initial illuminance data.

The lighting control system may further comprise a measurement system.

The proposed lighting control system may be an integral part of an integrated daylight and electric light control system and may further comprise an electronic model of the space being controlled, such as a Radiance model.

Fig. 2 illustrates an overview of a lighting control system comprising photo sensors 210, electric lights 220, blinds 230 and a controller 240. The photo sensors measure illuminance falling on the sensor locations. Illuminance data from the photo sensors 210 are used to carry out illuminance estimation 201 of the task lighting 203 through methods described in detail below. The controller 240 receives a setpoint 202 and the data from the illuminance estimation 201 and uses this to control the lights 220 and the blinds 230 in order to obtain a desired task lighting 203.

Fig. 3 illustrates an overview of a lighting control system comprising photo sensors 310, electric lights 320, blinds 330 and a controller 340. The photo sensors measure illuminance falling on the sensor locations. Illuminance data from the photo sensors 310 are used to carry out illuminance estimation 301 of the task lighting 303 through methods described in detail below. The light controller 340 receives a setpoint 302 and the data from the illuminance estimation 301 and uses this to control the lights 320. A separate blinds controller 350 controls the blinds 330. A desired task lighting 303 is obtained.

The received illuminance data may be classified into different categories based on environmental conditions such as exterior illuminance E_w, blind heights h and blind angles co. We may assume that the data are classified with the thresholds of exterior illuminance E , E₂,^■■■ , E_K_₁, the thresholds of blind height h , h₂,^■■■ , h_L_ and the thresholds of blind angle ω_1; ω₂,^■■■ , ω_Μ--1. Moreover, we may denote the minimum and maximum possible values of exterior illuminance as E₀ and E_K, h₀ and h_L for blinds height, and ω₀ and ω_Μ for blinds angle. Usually, E₀ = 0 and h₀ = 0 for positive exterior illuminance and blinds height. E_K and h_L may be two very large numbers compared with the actual exterior illuminance and blinds height. If the dynamic range of the blinds angle is from—90^° to 90^°, then ω₀ =—90^° and ω_Μ = 90^°. Thus data may be classified into K clusters by using the thresholds of exterior illuminance alone, into L clusters with the thresholds of blind height alone, and into M clusters for the thresholds of blind angle only. The maximum possible number of clusters may be K x L x M if the data are classified based on the combination of exterior illuminance, blinds height and blinds angle. However, it may be possible to have no data measured for certain combinations of these conditions. Therefore, the final number of data clusters may be less than or equal to K x L x M.

The detailed flowcharts of the data clustering are illustrated in Figures 4-6. These figures are provided as examples only. Here, we may assume that we first cluster for exterior illuminance, followed with clustering for blind height, and finally clustering for blind angle.

Fig. 4 illustrates a detailed flowchart of a first level data clustering for exterior illuminance. The first level data clustering loop comprises an initiating step 401 where a counter k is set to zero and the thresholds for exterior illuminance are predetermined as E₀, Ει_, ··· , E_K. In a controlling step 402 a check is made whether data has been measured in the first threshold range E_k < E_w < E_k+1. If data from this threshold range have not been measured, a check 403 is made whether the counter k is smaller than K-l. If this is the case, an incrementing step 404 increments to the next threshold range, whereby the controlling step 402 is performed again for the next threshold range. If the counter k is not smaller than K-l, the maximum threshold level has been reached and a terminating step 405 terminates the increment. If in the controlling step 402 there has been data measured in the first threshold range E_k < E_w < E_k+1, the counter k is incremented in step 406 and a check is made in step 407 to determine if the counter k has reached the end of this threshold range. If this is not the case, the controlling step 402 is performed again for this threshold range. If the counter k has reached the final threshold range, an obtaining step 408 obtains a data cluster of exterior illuminance E_k < E_w < E_k+1 for 0 < k < K— 1.

Fig. 5 illustrates a detailed flowchart of a second level data clustering for blind height. The second level data clustering loop is performed for each k obtained from the first level data clustering where 0 < k < K— 1 and comprises an initiating step 501 where a counter / is set to zero and the thresholds for blind height are predetermined as h₀, h_lt ^■■■ , h_L. In a controlling step 502 a check is made whether data cluster k has blind height data in the first threshold range h₁ < h < h_l+1. If data from this threshold range have not been measured, a check in step 503 is made whether the counter / is smaller than L-l. If this is the case, an incrementing step 504 increments to the next threshold range, whereby the controlling step 502 is performed again for the next threshold range. If the counter / is not smaller than L-l, the maximum threshold level has been reached and a terminating step 505 terminates the increment. If in the controlling step 502 there has been data measured in the first threshold range h₁ < h < h_l+1, the counter / is incremented in step 506 and a check in step 507 is made to determine if the counter / has reached the end of this threshold range. If this is not the case, the controlling step 502 is performed again for this threshold range. If the counter / has reached the final threshold range, an obtaining step 508 obtains an exterior illuminance and blind height data cluster (k, I) with E_k < E_w < E_k+1 and h₁ < h < h_l+1 for 0 < l < L - l.

Fig. 6 illustrates a detailed flowchart of a third level data clustering for blind angle. The third level data clustering loop is performed for each (k, I) obtained from the first and second level data clustering where 0 < k < K— 1 and 0 < I < L— 1 and comprises an initiating step 601 where a counter m is set to zero and the thresholds for blind angle are predetermined as ω₀, ω_1; ··· , ω_Μ. In a controlling step 602 a check is made whether data cluster (k, I) has blind angle data in the first threshold rangeeo^ < ω < o½₊₁. If data from this threshold range have not been measured, a check in step 603 is made whether the counter m is smaller than M-l. If this is the case, an incrementing step 604 increments to the next threshold range, whereby the controlling step 602 is performed again for the next threshold range. If the counter m is not smaller than M-l, the maximum threshold level has been reached and a terminating step 605 terminates the increment. If in the controlling step 602 there has been data measured in the first threshold range ω_τη < ω < o½₊₁, the counter m is incremented in step 606 and a check in step 607 is made to determine if the counter m has reached the end of this threshold range. If this is not the case, the controlling step 602 is performed again for this threshold range. If the counter m has reached the final threshold range, an obtaining step 608 obtains an exterior illuminance, blind height, and blind angle data cluster (k, I, m) with E_k≤ E_w < E_k+1 , h₁ < h < h_l+1 , w_m < ω < ½₊₁ for 0 < m < M - l.

In different embodiments of the present invention, either only one of these clusters or combinations of these clusters with different clustering order may be deployed, e.g. first starting with blind height, then exterior illuminance, etc. When roller shades are deployed, then the clustering may have only two dimensions: blind height and exterior illuminance. Clustering for only blind height or shade height might be the simplest strategy with respectable performance. Table 1 shows an example of the data classification with K = 3, L = 2, M = 6, in which boxes with integer numbers indicate that data are measured for the combined conditions of exterior illuminance, blind height and blind angle, while the remaining empty boxes indicate that no data are measured for those conditions. Firstly, data are clustered by the exterior illuminance into three groups {1,2,3, ·· · ,7}, {8,9,10, ·· · ,13} and {14,15,16}. Secondly, blind height is used to further cluster data into groups {1,2,3, ·· · ,6}, {7},

{8,9,10, ·· · ,12}, {13}, and {14,15,16}. Eventually, the data will be clustered into finer groups with individual numbers by further decomposing with blind angle.

Table 1: An example of data classification based on exterior illuminance, blind heights and blind angles, in which K = 3, L = 2, M = 6

Fig. 7 illustrates a method according to an embodiment of the present invention, comprising receiving 70 initial illuminance data. The initial illuminance data may be obtained from a data logger, wireless sensors, simulation models, etc. The method may comprise a classifying step 71 where illuminance data is classified into different categories using the clustering algorithms described above in relation to Figs. 4-6. Estimators may be calculated in a step 72, followed by a step 73 where measured illuminance data may be received from the at least one photo sensor 1 10. The taskplane illuminance may be estimated in a step 74. A detailed description of the calculation of the estimators and the estimation of the taskplane illuminance (step 72 and 74) is following below.

After clustering data into various conditions, the data may be investigated and estimators constructed separately for different conditions. Under the condition n, the workplane illuminance and the sensor illuminance may be measured and an optimal estimator may be constructed. This estimator may later be used to estimate the workplane illuminance from the surrounding sensors for the condition n.

Radiosity theory is used for the estimation of the workplane illuminance. For details on radiosity theory, see reference: Cindy M. Goral, Kenneth E. Torrance, Donald P. Greenberg and Bennett Battaile, "Modeling the interaction of light between diffuse surfaces ", Computer Graphics, vol. 18, no. 3, pp. 213-222, July 1984.

The daylight and electrical lights may be considered as a number of dimmable light sources P_j, for j = 1, 2, ·· · , J, and P ≥ P₂≥ ·· ·≥ P_j. Based on radiosity theory, the illuminance /_s on any surface in a room may be expressed as Is — ^aslPl + ^as2^2 ^"I^{" "}I^{" a}sjPj where _sl ... are constants. If the surface illuminance is measured at Q points, i.e.

>_> "' >IQ> ^{men a} matrix form expression of the surface illuminance vector /_s =

T

[k ·" IQ] is

Usually J > Q, and the modeled light sources P_Q+1,— ,Pj may be ignored. Therefore, an approximated expression of the surface illuminance is

When the matrix G is invertible, the modelled dimmable light sources are

P = G^_1/_s.

Based on the radiosity theory, the workplane illuminance may be expressed as a linear combination of the light sources

Itas_k = PlPl + P2P2 + - + P_jP_j ~ PlPl + P2P2 + - + P_QPQ = I^TsG-^TW₁ β₂ " _Q]^T where /?£... are constants.

Therefore, the workplane illuminance may be modeled as a linear combination of the measured surface illuminances has* = flk + h + - + fo = *^Tsf

For the discrete-time workplane illuminance signal

T

I task = [itaskil] Itaskl2]— haskW [ , the matrix notation of the workplane estimation is

Denote

then

^tasfe ⁼ Hf

This is a classic linear estimation (or regression) problem. The linear minimum mean squared error (LMMSE), i.e. the ordinary least squares (OLS), may provide a closed-form solution of the unknown coefficients / to minimize the sum of squared estimation error. = (^H) ¹^/ task

In some cases, may contain negative coefficients which are non-physical in reality. Therefore, the Tikhonov regularization (also known as Ridge regression) may be used to calculate a sub-optimal solution of coefficients = (H^TH + ry^t-ifi task where Γ is the Tikhonov matrix, which should be carefully selected to avoid negative coefficients. A most popular choice is the identity matrix Γ = I.

Furthermore, we may use an extra constant f₀ to represent the environment illuminance of the ignored light sources P_Q+1, - - - , P_j, which is

where 1 = [1 1 ··· 1]^T . Eventually, the estimation of workplane illuminance is expressed by hask = f₀ + flh + f2h + - + f_Q I_Q As mentioned previously, data may be clustered into various conditions. The same structure of linear estimation may be used for different conditions, but the estimator coefficients / may be discrete functions of the condition n. Thus, a general expression of the workplane illuminance estimation may be

The coefficients may be constructed by using the LMMSE

Or sub-optimal coefficients may be constructed to avoid large negative coefficients by using the Tikhonov regularization f[n] = (H[n]^TH[n] + r[n]^Tr[n])-¹H[n]^T/_tasJn]

And the constant may be f₀ [n] = l^T (Itask [n] - H[n]f[n])

It should be noted that the estimated coefficients may be calibrated automatically during the operation phase of the system to account for changing interior conditions and different seasons. However, is should be noted that this auto calibration is not necessary for the present invention.

Measured illuminance data may be received in a step 73 from the at least one photo sensor 110. An illuminance setpoint may be received in a step 75. The controller may provide 76 control signals to the electric light sources and the shading system. The lighting control system may be operated in such a way that the steps 70-76 are repeated until the difference between the user setpoint and the estimated taskplane illuminance has been acceptably minimized. Thus the desired amount of light at the workplane may be obtained. It may be noted that the shading system 130 may be controlled using different open- loop strategies such as those based on sun angle, solar clock with external sensors, and detailed model of the building and surroundings. In addition to the method described above, in relation to Figs. 1-7, it may also be possible to perform a complementary method that will be described in relation to Fig. 8. It should be noted that this complementary method may also be performed separately. It may also be noted that this complementary method may be utilized as an add-on to a lighting optimization algorithm. The complementary method described below in relation to Fig. 8 is an illuminance model generation method and may be used for inferring daylight contribution on each point of interest in a daylit space. Performing of this method may effectively separate workplane illuminance contributed by electric lights from illuminance contributed by daylight in a daylit space. The knowledge obtained from the illuminance model may be used for control purposes, such as blind control.

Fig. 8 illustrates a method according to an embodiment of the present invention, comprising turning off 80 a dimmable electric light source. Initial illuminance data may be received in a step 81 from photo sensors. The method may comprise an actuating step 82 where the dimmable electric light source is actuated to a prescribed level. Illuminance data is again received from the photo sensors in a step 83. An illuminance model is derived in a step 84. The electric light source contribution may be calculated in a step 85. The steps 80- 85 will be described in detail below. The method in Fig. 8 further comprises a step 86 where an illuminance setpoint may be received. The controller may provide 87 control signals to the electric light sources and the shading system in order to generate a desired task illuminance at the point of interests whilst the usage of the exterior light contribution is optimized.

The steps 80-85 in Fig. 8 relate to a procedure that may generate a workplane illuminance model for a set of electric light sources or luminaires. This model generation may be built on an infrastructure comprising at least one photo sensor 110, at least one electric light source 120 and a lighting actuator 150 which may actuate the at least one electric light source 120 to different dimming levels. The procedure might run for every electric light source or luminaire while it is being dimmed to different levels. The task illuminance may be measured for each photosensor and may be associated with the corresponding luminaire and its output level. This forms the illuminance model. The method may further separate the workplane illuminance contributed by electric lights from that contributed by daylight in a daylit space. Due to the linearity of the illuminance models, the control system may calculate the exact electric light contribution, thereby separating it from the daylight contribution on a particular point of interest in the space being controlled. This knowledge of net daylight on each point of interest may then be used for the control of a shading system. It may be noted that this procedure preferably runs during off hours when there is no or limited daylight.

It may be assumed that there are K luminaires. It may be assumed that there are N points of interest in the space to be controlled. A point of interest may be the surface of a desktop. It maybe assumed that V (d) = [l (d) ^■■■ V_N (d)J represents the workplane illuminance model as a function of an actuation level d. The actuation level d may represent a prescribed lighting level of a dimmable luminaire. The workplane illuminance model may be automatically generated for the i^th luminaire by utilizing Nworkp lane-based photosensors deployed on each point of interest. In order to obtain l[ {d) , the lighting control system may follow the following steps for each luminaire:

Turn off the i^th luminaire;

Receive and record the initial sensor readings from each point of interest from workp lane-based photosensors, denoted as {e_j' (θ )} _t ;

Actuate the luminaire to a prescribed level d;

Receive and record the sensor readings from each point of interest, represented as {e) {d )}^N _j=i ;

Derive the illuminance model for luminaire i at each points of interest of j as j {d ) = e_j ⁱ {d ) - e_j ⁱ {Q )Y_j=i ,-

Actuate the luminaire to the next prescribed level;

Repeat step 4 to 6 until all prescribed actuation levels are measured.

Further details of the workplane- level illuminance models that are automatically generated above and their contribution to the lighting control system are presented below. Denote ei, e_# as the illuminances generated by the electric lights at each point. Elements in a vector Eo are the illuminances contributed by daylight at each of the N workstations, d = [di, d_k]^T are the actuation levels, i.e. the light output settings, of the K luminaires. The relationship between the illuminances at the points of interest, the illuminance models, and the luminaire actuation levels may be represented in the following matrix operation.

E₀ may be derived from E₀ = E(d) - L(d), where elements of E(d) are measured with the workstation-based photosensors and L(d) may be calculated from the illuminance models given the current luminaire actuation levels d. The knowledge of net daylight contribution, E₀, may then be used for the control of blind system. In order to find a new optimal set of luminaire actuation levels (d) to result in desired illuminances at each point of interest ( desired) given current daylight condition E₀, the lighting controller may formulate the lighting actuation problem into a linear programming problem as shown below. d - min d ,

d

subject to

E(d)e E_desired ,

DimLevel ■ < d < DimLevel„

The objective function of the linear programming problem minimizes the 1- norm of the column vector d, which is the summation of each element in d. In other words, the summation of luminaire light outputs is minimized. Since the light output from a luminaire is proportional to its energy usage, minimizing the overall light output equals minimizing energy consumption. In the meantime, the linear programming problem subjects to two sets of constraints. The first set of constraints (E(d)e desired) specifies that the resulting lighting at each point of interest must be within the desired range of illuminance. The second set of constraints (DimLevel_mi„ < d < DimLevel_max) accounts for the physical limitations of the luminaires. The lower and upper bounds of the inequality denote the minimum and maximum possible output of the luminaires respectively. Consequently, the solution to the linear programming problem (d^*) is the set of luminaire actuation settings that minimizes energy usage while still delivering specified lighting to each point of interest.

Illuminance data of two ceiling sensors and one workplane sensor were measured in an office room on August 17^th, 2010 to construct estimators, which were used on August 20^th, 2010 to estimate the workplane illuminance from the other two ceiling sensors.

Figs. 9a and 9b show the comparison of the measured workplane illuminance and the estimated workplane illuminance without data clustering (Fig. 9a) and with data clustering with 14 clusters (Fig. 9b) for a sunny day on August 20^th, 2010. Similarly, Figs. 9c and 9d are the comparisons of performance for a cloudy day on August 21^st and Figs. 9e and 9f for a rainy day on August 22^nd, respectively. Apparently, the data clustering improves the estimation performance for different weather conditions. It should be noted that only one clear day (August 17^th) was used to construct estimators.

The person skilled in the art realizes that the present invention by no means is limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims.

Claims

CLAIMS:

1. A method for operating a lighting control system (100) comprising at least one photo sensor (110), at least one dimmable electric light source (120), a motorized shading system (130) and a controller (140), the method comprising:

receiving initial illuminance data based on at least two environmental conditions n, wherein said at least two environmental conditions n are chosen from the group of exterior illuminance Ew, blind height h, and blind angle ω;

classifying said initial illuminance data based on predetermined thresholds of said at least two environmental conditions;

calculating estimators for said at least two environmental conditions n;

receiving measured illuminance data from at least one photo sensor (110); estimating a taskplane illuminance I_task for a current set of environmental conditions n by determining said taskplane illuminance I_task as a function of said calculated estimators and said measured illuminance data;

receiving a taskplane illuminance setpoint;

providing control signals to said at least one dimmable electric light source (120) and said motorized shading system (130), thereby minimizing the difference between said taskplane illuminance setpoint and said estimated taskplane illuminance I_tas

2. The method according to claim 1, wherein said classifying step comprises classifying data into K clusters by using said predetermined thresholds of said exterior illuminance E_lt E₂, ··· , E_K_₁, wherein the minimum and maximum possible value of said exterior illuminance is denoted as E₀ and E_K respectively.

3. The method according to claim 1 or 2, wherein said classifying step further comprises classifying data into L clusters by using the thresholds of blind height

h-L, h₂,^■■■ , h_L_ , wherein the minimum and maximum possible value of blind height is denoted as h₀ and h_L respectively.

4. The method according to claim 3, further comprising classifying data into M clusters by using the thresholds of blind angle ω_1; ω₂,^■■■ , ω_Μ-ι, wherein the minimum and maximum possible value of blind angle is denoted as ω₀ and ω_Μ respectively.

5. The method according to claim 1 , wherein said controller (140) is comprising a light controller (340) and a blinds controller (350) and wherein said providing control signals step comprises providing control signals to said at least one dimmable electric light source (120, 320) from said light controller (340) and providing control signals to said motorized shading system (130, 330) from said blinds controller (350).

6. The method according to claim 1, wherein said estimated taskplane illuminance I_task is determined according to hask ⁼ sl (ⁿ) l + ^■■■ + fsm hm + fs(m+l) ( )^s(m+l) +^■■■ wherein f_sm is an unknown estimator coefficient pertaining to a weighting scale of an m^th photo sensor during said at least one environmental condition n and l_sm is the illuminance value measured by said m^th photo sensor.

7. The method according to claim 6, wherein said estimated taskplane illuminance I_task is a linear combination of measured illuminance values and is modeled according to

wherein I_{L T} I₂ , · · · , IQ are illuminance values measured at Q points from a total number of Jpoints, wherein J > Q, thereby ignoring the illuminance values IQ ₊₁, ·· · , I_j, and wherein f₁, f₂, · · · ,†Q are unknown estimator coefficients.

8. The method according to claim 7, wherein said unknown estimator coefficients are constructed by using least minimum mean squared error (LMMSE) and are estimated according to = (ifHrWltask wherein H and I_taSk ^are denoted in matrix form as

9. The method according to claim 8, wherein Tikhonov regularization is used to calculate said unknown coefficients according to

J = {H^TH + r^Try¹H^Ti_task wherein Γ is the Tikhonov matrix being selected to avoid negative coefficients.

10. The method according to claim 7, wherein said estimated task lane illuminance I_task is modeled according to

wherein ₀is an extra constant representing the ignored illuminance values wherein f₀ is determined according to

wherein 1 = [1 1 ·· · 1]^T .

11. The method according to claim 1, wherein said at least one photo sensor (110) is at least one taskplane photo sensor located at at least one point of interest in a taskplane and wherein said lighting control system (100) is further comprising a lighting actuator (150) for said at least one dimmable electric light source (120), the method further comprising the steps of

turning off said at least one dimmable electric light source (120); receiving initial illuminance data from said at least one taskplane photo sensor; actuating said at least one dimmable electric light source (120) to at least one prescribed lighting level;

receiving illuminance data from said at least one taskplane photo sensor;

deriving, for said at least one point of interest, an illuminance model for said at least one dimmable electric light source (120) being actuated to said at least one prescribed lighting level;

calculating an electric light illuminance contribution, thereby separating said electric light illuminance contribution from an exterior illuminance contribution, for said at least one point of interest;

receiving a desired illuminance setpoint for said at least one point of interest; providing control signals to said at least one dimmable electric light source (120) and said motorized shading system (130), thereby generating a desired task illuminance at said at least one point of interest.

12. A lighting control system (100) comprising:

at least one photo sensor (110),

at least one dimmable electric light source (120),

a motorized shading system (130), and

a controller (140),

wherein said controller (140) is adapted to

receive initial illuminance data based on at least two environmental conditions n, wherein said at least two environmental conditions n are chosen from the group of exterior illuminance Ew, blind height h, and blind angle ω;

classify said initial illuminance data based on predetermined thresholds of said at least two environmental conditions n;

calculate estimators for said at least two environmental conditions n;

receive measured illuminance data from said at least one photo sensor (110); estimate a taskplane illuminance I_task for a current set of environmental conditions n by determining said taskplane illuminance I_task as a function of said calculated estimators and said measured illuminance data;

receive a taskplane illuminance setpoint;

provide control signals to said at least one dimmable electric light source (120) and said motorized shading system (130), thereby minimizing the difference between said taskplane illuminance setpoint and said estimated taskplane illuminance.

13 The lighting control system according to claim 12, wherein said controller

(140) comprises a light controller (340) and a blinds controller (350), and wherein said controller is further adapted to provide control signals to said at least one dimmable electric light source (120, 320) from said light controller (340) and to provide control signals to said motorized shading system (130, 330) from said blinds controller (350).

14. The lighting control system according to claim 12, further comprising a taskplane illuminance sensor being used for measuring said initial illuminance data.

15. The lighting control system according to claim 12, wherein said at least one photo sensor (1 10) is at least one taskplane photo sensor located at at least one point of interest in a taskplane, wherein said system is further comprising a lighting actuator (150) for said at least one dimmable electric light source (120), wherein said controller (140) is further adapted to

turn off said at least one dimmable electric light source (120); receive initial illuminance data from said at least one taskplane photo sensor

(1 10);

wherein said actuator (150) is adapted to

actuate said at least one dimmable electric light source (120) to at least one prescribed lighting level;

wherein said controller (140) is further adapted to

receive measured illuminance data from said at least one taskplane photo sensor;

derive, for said at least one point of interest, an illuminance model for said at least one dimmable electric light source (120) being actuated to said at least one prescribed lighting level; calculate an electric light illuminance contribution, thereby separating said electric light illuminance contribution from an exterior illuminance contribution, for said at least one point of interest;

receive a desired illuminance setpoint for said at least one point of interest; provide control signals to said at least one dimmable electric light source (120) and said motorized shading system (130), thereby generating a desired task illuminance at said at least one point of interest.