Integrated path planning for a partially unknown outdoor environment

In this paper, we propose an integrated path planning algorithm that reasons on integration between a local path planners based on sensory information and global path planners based on global prior information in a partially unknown outdoor environment. A local path planner typically generates collision-free paths using laser range sensors, sonar sensors, etc., nearby robots or unmanned ground vehicles. In contrast, a global path planner generates efficient paths based on given cost function using initially given maps. To generate efficient collision-free paths in a partially unknown outdoor environment, we utilize both information properly in RRT as the local path planner and A* as global path planner and separate a role of each planner not to overlap with each other. It is assume that a partially unknown outdoor environment only provides obstacle information nearby a robot or a vehicle within a range of a sensor and prior approximate information about the entire map. The proposed integrated path planner improves overall computational time and cost to obtain one of the complete path from start to goal.