Cellular neural network-based thermal modelling for real-time robotic path planning

Abstract

This paper presents a heat conduction method for optimal robot path planning by drawing an analogy between heat conduction and path planning of mobile robots. This method treats the target as a heat source to propagate the target activity through heat conduction. Obstacles in the configuration space are treated with Dirichlet boundary conditions such that the resultant heat flux automatically causes the robot to avoid them. A transient-state heat conduction model is established to generate the temperature field in the state space, guaranteeing that the target and obstacles remain at the peak and the bottom of the heat distribution, respectively. A cellular neural network is also established for the real-time modelling of the heat conduction process. An iterative heat flux method is established to generate an optimal path from the robot to the target. The proposed method can not only generate real-time, optimal and collision-free paths without any prior knowledge of the dynamic environment and without explicitly searching over the global free work space or searching collision paths, but it can also easily respond to the real-time changes in dynamic environments.