基于单神经元自适应PSD算法的分布式智能灯光控制

Distributed smart lighting control based on single neuron adaptive PSD

针对已有灯光控制算法无法应对系统模型变化的问题,提出将单神经元自适应PSD(proportion sum differential)算法应用于分布式智能灯光控制。利用单神经元自适应PSD算法的自学习能力,控制器根据系统误差实时修改参数,并与无线传感器/执行器网络中的分簇机制相结合,形成了一套完整的自适应分布式智能灯光控制算法。以基于无线传感器/执行器网络的灯光控制实验平台为被控对象,设计控制器并进行了仿真研究。仿真实验表明,当系统模型发生改变,与已有的分布式PID灯光控制算法相比,该控制算法具有更好的控制效果、鲁棒性更强。

This paper developed a new distributed adaptive smart lighting control algorithm,which based on single neuron adaptive proportional sum differential（ PSD） algorithm and clustering mechanism of wireless sensor and actuator networks.Due to single neuron adaptive PSD had the ability of self-learning and could modify control parameters online based on system errors,the new controller could be used in a complex environment such as system model changing. It designed a controller for the experiment platform of lighting control based on WSANs,simulation shows that compared to distributed PID lighting control system,the new controller are more robust and can achieve good control results even when system model changes a lot.