Space deformation based path planning for Mobile Robots

An excellent path planning algorithm of a robot should compromise three major criteria; low computational time, high level of smoothness and optimal length. In this work, a hybrid algorithm is developed to enable the robot to navigate smoothly in a partially known environment with a low computation time. The proposed method takes as input a global path connecting a start and a target point, then an initial optimal smoothed path is generated which is accordingly updated due to unexpected changes in the workspace. This is achieved by assuming that the robot exits on a thin metal plate which can be deformed to guarantee a convenient path for the robot. This deformation of the space continues as long as changes are detected in the environment. Numerical simulations and experimental analysis showed the high performance of the proposed algorithm, where it showed superiority in terms of smoothness and execution time.