The active kinematic histogram method for path planning of non-point, non-holonomically constrainted mobile robots

The paper presents a novel path planning method for non-point, non-holonomically constrainted mobile robots. The proposed method, named the Active Kinematic Histogram (AKH) method, is a potential field path-planning method based on the Vector Field Histogram (VFH) method introduced by J. Borenstein and Y. Koren. The method relies on one-dimensional environment representations named AKHs, derived from the available information on the obstacle space and the robot itself. The AKHs provide the information abstraction required for fast, collision-free and kinematically feasible path planning for non-point, non-holonomically constrainted mobile robots. The method constructs two AKH structures incorporating the information stored in occupancy grid environment representations as well as the exact robot's size, shape and kinematic features. The method applies simple and fast procedures, e.g. histogram shifting, in order to determine kinematic feasibility and the result of the robot's steering limits. Being a potential field method, the AKH method uses adaptive thresholding procedures to reduce the inherited drawback of the local minima traps phenomenon. Furthermore, the method proposes specific updating and coding procedures for the environmental information and introduces the Ageing function to insert time parameterization in the path-planning results. Also, the so-called bicycle kinematic model is presented in order to describe the kinematic family of mobile robots suitable for applying the AKH method. The method integrates a velocity control procedure which is also based on information provided by the active kinematic histograms. The method produced smooth, collision-free paths in a variety of experimental configurations. Simulation and experimental results are presented.