Intelligent Lighting Control Based On Wireless Sensor And Actuator Networks

With the development of social economy and accelerating process of urbanization,the ratio of lighting electricity in total power consumption has been increasing. While intelligent lighting system which is the important part of smart building has attracted extensive attention worldwide and developed rapidly. Meeting users’ personalized requirements of lighting with the least energy consumption is the main target of intelligent lighting control system. Wireless sensor and actuator networks(WSANs) composed of sensor and actuator nodes not only can sense the surrounding environment by sensor nodes but also can change environment with the help of actuator nodes. Communications between nodes are in a wireless way so are easy to install. Intelligent lighting control is one of important application fields of WSANs. This paper research on an intelligent lighting control experiment platform based on WSANs, and the main work is concluded as follows:1. In order to improve the accuracy of occupancy detection and location, replaced passive infrared occupancy sensors used before in the lighting control experiment platform with a camera node. The new occupancy detection system is based on single view metrology and foreground-background segmentation with Codebook background model. Proposed a real time human detection and location method based on single camera and applied in lighting control system, verified its feasibility and validity through experiments.2. Combined Particle Swarm Optimization(PSO) and PID neural network(PIDNN) algorithm and used to control lighting which is a MIMO system. Designed a controller based on PSO-PIDNN algorithm for the experiment platform. The performance of controller is researched by simulation.3. The existing control methods of smart lighting based on WSANs are most depending on the accurate model of controlled system, and can’t deal with the mutation of system. Developed a distributed lighting control method based on single neuron adaptive PSD algorithm and coordination mechanism of actuator nodes in cluster. The controller has the ability of self-learning and can modify parameters online based on errors so is more robust. Designed distributed controllers for the experiment platform and simulation verified that controllers could have good control results even when system model changes a lot. The control algorithm is easy to implement, robust and convenient to expand, so has a high applicable value.