Memory-time tradeoffs in a path planning approach utilising limited memory robots

Abstract

Mapping a large environment using a group of heterogeneous mobile robots can be problematic when some (or all) of the robots do not have sufficient memory to store the entire global map at the required resolution. However, in certain applications, these memory constrained robots may need to perform global path planning to navigate beyond their local region. A two-tiered path planning technique based on the A* algorithm is presented to facilitate successful global path planning. The technique involves initially dividing a large global map into smaller local maps whose size is determined by the robot’s memory capacity. Following this, a path to the destination is planned by searching the local maps using a two-tiered A* algorithm. The effect of local map size on the path length and planning time is investigated for different global map sizes and obstacle densities.