Complete coverage path planning (CCPP, also called region filling or area covering) is a special type of trajectory generation that requires the robot path to pass through the every area of the workspace. Many other robotic applications such as painter robots, land mine detectors, lawn mowers, and window cleaners also require CCPP. In this thesis, a biologically motivated neural network strategy is proposed for CCPP of cleaning robots with obstacle avoidance in changing environments. The robot coverage motion is generated through the dynamic activity landscape and neural activity propagation of the neural network. There are only local connections in the topologically organized neural network. No templates, no prior knowledge of the dynamic environment, and no learning procedures are needed. The global stability and convergence of the neural network system are guaranteed by Lyapunov theory. The model can be applied to a single robot and multiple robots. The planned robot paths with designated starting and goal locations do not suffer from deadlock problems.
展开▼
