Intelligent Motion Planning of a Collaborative Mobile Robot in a Dynamic Environment Based on VFH+ and Semantic Processing of LIDAR and Video Data

Abstract

The article presents results of the study aimed at developing and substantiating a method for intelligent motionplanning of a collaborative mobile robot in a dynamic environment based on the VFH+ algorithm with semantic weighting of LiDAR and video camera data. The purpose of the study is to enhance the safety, predictability, and adaptability of a mobile robotic platform operating in the presence of static and dynamic obstacles through the integration of geometric environmental information with contextual risk assessment. The object of the study is the process of planning and executing the trajectory of a collaborative mobile robot within a structured workspace that includes dynamic elements typical of Human–Robot Collaboration (HRC) scenarios. The subject of the study comprises mathematical models and algorithmic procedures for generating control actions based on a polar obstacle histogram augmented with a semantic evaluation of the hazards associated with objects detected through computer vision. The study uses a kinematic model of a differential-drive system, a method for constructing a polar obstacle grid using LiDAR data, generation and smoothing of the VFH+ histogram, sector binarization and identification of admissible motion directions, as well as a multi-criteria selection of the optimal sector considering goal orientation, smoothness, and risk level. To consider the context of interaction with humans and other objects, a semantic multiplier is introduced, which modifies the initial histogram according to the results of video detection. The numerical modeling is implemented in Python with discrete integration of motion equations, taking into account sensor noise and combining local VFH+ with global route landmarks obtained with the use of A* algorithm.