面向室外环境的单机器人辐射场寻源算法

针对室外热辐射场、核辐射场、强电磁辐射场和声场的辐射源搜寻,提出了一种基于高斯过程回归的单机器人辐射源搜寻算法。首先将连续的搜寻环境离散化,采用高斯过程回归预测机器人邻域中的辐射强度,然后将局部路径规划考虑为一个优化问题,求取优化目标函数的最大值获得机器人的下一步路径点,最后通过多次局部路径规划引导机器人到达全局最优点。为验证所提算法的有效性,在300 m×200 m的室外环境中进行了实验。

A Radiation Source Searching Algorithm Used for a Single Robot in Outdoor Environment

Aiming at searching the source of thermal radiation fields, nuclear radiation fields, strong electromagnetic radiation fields and sound fields in outdoor environment, a source searching algorithm based on Gaussian Process Regression using a single robot is proposed. First of all, the continuous searching area is discretized with square grids, and the radiation intensity of adjacent grids is predicted by Gaussian Process Regression. Then a local path planning is considered to be an optimization problem and the next path point is obtained by solving the problem. Finally, the robot is guided to the global optimal point through repeated local path planning. To verify the proposed algorithm, an experiment in an environment of 300 m×200 m has been conducted.