A Reinforcement Learning (RL)-Based Hybrid Search Method for Hidden Object Discovery using GPR

Ground Penetrating Radar (GPR) systems are effective tools for discovering various types of buried objects (such as military mines, metal objects, pieces of underground infrastructures etc.). One challenge is to find an efficient search procedure that would reduce the operation time and optimize resource utilization, such as minimizing the power consumption for battery operated UAVs or fuel consumption of land vehicles. Current approaches are based on pre-defined search patterns which, for large and sparse areas, could mean unnecessary waste of time and resources. In this paper we report on a Reinforcement Learning (RL) based search method for training the guidance logic of possible platforms hosting the GPR, so that they navigate more efficiently. We illustrate the applicability of the method using a high-fidelity simulation environment and a newly implemented RL framework. Initial results suggests that combination of a coarse navigation scheme and an RL based training procedure based on GPR scan patterns, can lead to a more efficient target discovery procedure for host platforms.