[0044] Such as figure 1 As shown, an open office distributed lighting control method, the open office includes:

[0045] N LED lights with adjustable brightness arranged on the ceiling, and M workbenches on a horizontal plane 0.8m above the ground

[0046] Smart devices arranged on M workbenches to set the expected illuminance of the workbench, the expected illuminance set by each workbench is broadcasted by the smart device in a wireless form;

[0047] N sensor-actuator nodes arranged on the ceiling are used to detect the illuminance of the LED lights in a one-to-one correspondence, and adjust the brightness level of the N luminaires in a distributed manner based on the deviation between the expected illuminance and the actual illuminance, and each LED light is based on the duty cycle The ratio mechanism controls and adjusts the brightness from 0% to 100%;

[0048] The present invention includes the following steps:

[0049] (1) Representation of graph theory:

[0050] Use graphs to characterize the communication relationship between nodes in the distributed network composed of various sensor-actuator nodes, where V={1,2,...,N} is the set of nodes, and N represents the number of nodes;

[0051] Edge set Represents the communication relationship between N nodes, if node i receives the information of node j, then {i,j}∈E(i,j∈V,i≠j), and vice versa And suppose there is no self-connection in the network, namely If {i,j}∈E, then node j is the input neighbor node of node i;

[0052] Define N i ={j|(i,j)∈E,i≠j} represents the set of input neighbor nodes of node i, d i =|N i | Represents the in-degree of node i.

[0053] (2) Calculate the expected illuminance of the open office based on the average consistency theory:

[0054] Remember ω rl Is the expected illuminance setting value of worktable l, where l=1, 2,...,M), ω rl It is set by the wireless smart device located on the workbench and broadcast to the sensor-actuator node in a wireless manner after the setting is completed; the sensor-actuator i located on the ceiling receives the expected illuminance setting value sent by the workbench within the communication range ω rq (q=1,2,...,M,q∈N ri ), according to the formula Perform initial office expected illuminance ω ri (0) is obtained by connecting with the adjacent sensor-actuator j(j∈N i ) Information exchange, and based on the formula Proceed ω ri (k) iteration, where k∈Z is the number of iterations, μ> 0 is distributed control gain, W ij Is the connection weight between sensor-actuator nodes j and i;

[0055] Use network G to characterize the communication relationship between N sensors and actuators, and record the diagonal array Is the in-degree matrix of network G, W={W ij } Is the connection weight matrix of network G, then the Laplacian matrix of network G is L=D in -W, respectively according to the formula:

[0056]

[0057] And the formula:

[0058]

[0059] Go W ij And μ design, where d i Is the number of input neighbor nodes of sensor-actuator node i, λ i (L) is the i-th eigenvalue of matrix L;

[0060] (3) Distributed adjustment of LED light brightness based on the deviation between expected illuminance and actual illuminance, including the following steps:

[0061] (3.1), daylight matching matrix C LS estimate:

[0062] Mark d i And p l Are the illuminance contribution of daylight at sensor-actuator i and workbench l(, d and p are the illuminance contribution vector of daylight at sensor-actuator and workbench respectively, where the illuminance contribution vector d is composed of N sensors- Measured by the actuator;

[0063] In the system training phase, turn off all LED lights, arrange M illuminance sensors on M workbenches to detect the illuminance contribution p of sunlight on the workbenches, measure r group data to obtain sample matrix P=[p 1 ,p 2 ,...,p r ] And D = [d 1 ,d 2 ,...,d r ], r> > M,r> > N, the daylight matching matrix C from d to p based on the least square method LS ∈R M×N The estimate of is specifically:

[0064]

[0065] In this way, in the lighting control system, according to the illuminance contribution d of the sunlight at the sensor-actuator, the illuminance contribution p of the sunlight at the workbench is estimated, specifically p=C LS d.

[0066] (3.2) Real-time illuminance calculation of workbench:

[0067] Mark y separately i , U i And ω l For the actual illuminance detected by the sensor-actuator i, the brightness output of the LED lamp i and the actual illuminance of the worktable l(, the sensor-actuator i is in accordance with the formula:

[0068]

[0069] Estimate the actual illuminance of the worktable l(, where: A ij , B lj They are the illuminance gain of LED light j at sensor-actuator i, the illuminance gain of LED light j at workbench l(, d i (k) and p l (k) are the illuminance contribution of sunlight at sensor-actuator i and workbench l(; Ξ ij =1 means that the sensor-actuator node i can receive the information of the sensor-actuator node j, and Ξ ij =0 means that sensor-actuator node i cannot receive the information of sensor-actuator node j, Γ lj =1 means that the sensor-actuator node j can receive the information at the workbench 1, Γ lj =0 means that the sensor-actuator node j can receive the information of the workbench l(, Indicates the matching gain from the contribution of sunlight at the sensor-actuator i to the contribution of sunlight at the workbench l;

[0070] (3.3) Distributed adjustment of LED brightness:

[0071] Sensor-actuator i follow Determine the workbench m most affected by the LED light i, and according to e m (k)=ω rm -ω m Calculate the deviation between the expected illuminance and the actual illuminance e at the worktable m m (k), according to the formula u i (k)=Q{u i (k-1)+α i Γ mi e m (k-1)) Carry out LED light i brightness output value u i The calculation of (k), where Q{.} is a function to ensure that the brightness output value is between [0,1], defined as:

[0072]

[0073] Where α i To control the gain, follow Perform distributed design.

[0074] Simulation results and analysis

[0075] In order to illustrate the effectiveness of the distributed lighting control method proposed in the present invention, simulation verification of the algorithm was performed on the DIALux platform and the Matlab platform. In a 30×15×2.8(m 3 ) The lighting system in the open office such as figure 2 As shown, where M=28, N=80; the communication topology diagram is obtained when the wireless communication radius is set to 6m. image 3 As shown, the expected illuminance error ω of the office at the sensor-actuator i (i=1,...,N) ei (k) Iterative curve such as Figure 4 As shown, the comparison curve of lighting energy consumption with the other two methods is as Figure 5 Shown. From image 3 with Figure 4 It can be seen that the office average illuminance estimation method proposed in the present invention is effective, and after 50 iterations of the average illuminance estimation algorithm, the expected illuminance error can be less than 10 -10 lux; from Figure 5 It can be seen that, compared with the illuminance non-control system and the illuminance reference control method, the method provided by the present invention has the lowest energy consumption, and the lighting energy consumption can take into account the influence of the sunlight.