A Radioactive Source Searching Algorithm Based on Gaussian Process Regression

10.6567/IFToMM.14TH.WC.PS13.002

Shi-Xiang Ni；Yan-Yu Su；Wei-Dong Wang；Dong-Mei Wu；Zhi-Jiang Du

Autonomous robot ； Gaussian Process Regression ； Radioactive sources ； Searching algorithm

Proceedings of the 14th IFToMM World Congress

14th-5（2015 / 11 / 06）

401 - 406

英文

Robots have been applied to search for radioactive sources. Until now, however, it is still an open problem to accomplish this task by one autonomous robot. In this paper, we propose an algorithm based on Gaussian Process Regression (GPR) to search for a radioactive source in a discrete 2D environment. We predict the radiation intensities of 8 positions using GPR. We consider searching a radioactive source as an optimization problem so as to obtain target positions for a mobile robot. The objective function describes three aspects: the radiation intensity, the amount of information for GPR, and the environment with obstacles. Simulation experiments are conducted to verify the proposed algorithm. The simulation results show that our algorithm is effective and robust when searching radioactive sources located at different points in an indoor environment.