Genetic algorithm based optimal placement of PIR sensor arrays for human localization

Abstract

The optimal pyroelectric infrared (PIR) sensor arrays placement similar to optimal multi-camera placement in computer vision (CV), aims to maximize the coverage and spatial resolution and minimize the cost. In this paper, we propose an implementation of a genetic algorithm (GA) based optimization approach for the design of PIR sensing model which is designed empirically in the literature. The optimization process entails the deployment of the sensors and the modulation of sensors' fields of view (FOV). The conventional design need more prior knowledge on the sensing system, while the proposed optimization approach enables the design more flexible and accurate with little prior knowledge. This optimization approach is illustrated by designing a PIR sensing model for human-locating system, and the experimental results testify the validity of the GA-based design approach.