Heat conduction combined grid-based optimization method for reconfigurable pavement sweeping robot path planning

Self-reconfigurable robots can change their morphology to expose functionality that enables them to overcome challenges that fixed shape robots are unable to overcome. A reconfigurable pavement sweeping robot, Panthera, is able to reconfigure in width by compressing and expanding. This autonomous system overcomes challenges in the pavement cleaning industry, such as access to tight and narrow pavements. Reconfigurability in width enables the robot to maximize cleaning area in an expanded state and navigate through tight spaces in the compressed state, which fixed shape robots are unable to pass through. Despite its advantages, most path planning (PP) algorithms are developed for fixed shape robots, and there are little or no works done on PP for reconfigurable robots that are able to expand and compress in width. To tackle this challenge, we present a novel PP method for pavement-sweeping reconfigurable robots by drawing similarities between heat transfer analysis and PP. Heat travels from the heat source to the heat sink and can only travel through a conductive material. Similarly, PP enables the robot to move from start to goal while accessing the feasible areas. Our proposed path planner employs a thermal gradient ascent combined grid-based optimization algorithm to generate optimal paths for best smoothness, distance, and robot footprint. The path planner is tested in four virtual environments and validated in a real-world environment.