Target Search and Navigation in Heterogeneous Robot Systems with Deep Reinforcement Learning
This addresses search and rescue efficiency in unknown, maze-like environments, but it is incremental as it builds on existing deep reinforcement learning methods with specific modifications.
The paper tackles the problem of target search and navigation in unknown environments using a heterogeneous robot system (UAV and UGV) for search and rescue, achieving successful training with a multi-stage reinforcement learning framework and curiosity module that accelerates training speed compared to baselines.
Collaborative heterogeneous robot systems can greatly improve the efficiency of target search and navigation tasks. In this paper, we design a heterogeneous robot system consisting of a UAV and a UGV for search and rescue missions in unknown environments. The system is able to search for targets and navigate to them in a maze-like mine environment with the policies learned through deep reinforcement learning algorithms. During the training process, if two robots are trained simultaneously, the rewards related to their collaboration may not be properly obtained. Hence, we introduce a multi-stage reinforcement learning framework and a curiosity module to encourage agents to explore unvisited environments. Experiments in simulation environments show that our framework can train the heterogeneous robot system to achieve the search and navigation with unknown target locations while existing baselines may not, and accelerate the training speed.