Q-learning-Based Path Planning with Terrain-Aware Concealment in Ground Battlefield Environments

Path planning in ground battlefield environments requires more than finding the shortest path, as safety and survivability are critical under uncertain and hostile conditions. Conventional path planning methods mainly focus on obstacle avoidance, which is insufficient for battlefield scenarios where terrain can be used for concealment and cover. This paper proposes a Q-learning-based path planning method that exploits terrain features to enable safe and tactical movement in ground battlefield environments. The environment is modeled as a two-dimensional grid map with various terrain types. A reward function is designed to consider both path efficiency and terrain-based concealment and cover, allowing the agent to learn strategic paths rather than purely shortest routes. Experimental results from multiple target scenarios show that the proposed method achieves high success rates and stable performance, even in complex environments. The results demonstrate that terrain-aware reinforcement learning is effective for path planning in battlefield environments and can support autonomous ground vehicles and military robotic systems.