Social robot navigation in human-robot interactive environments: Social force model approach

Abstract

The aim of this paper is to discover strategies focusing on maintaining humans' physical safety and mental comfort in human-robot interactive social environments. Our work builds on the ideas for attaining human-like collision avoidance behavior for mobile robots that share the same social environment with people. More explicitly, this paper intends to find out a social motion planning algorithm that enables the robot to have from “regular collision avoidance” to “human-like collision avoidance”. To achieve this goal, we use a variant of a pedestrian model called Collision Prediction based Social Force model that is particularly developed for low or average density environments and also benefits from human motion during navigation. Once the model parameters have been calibrated, the model is then adopted to be used as a local planner to generate a socially aware path planning in a hallway scenario. The system is tested in simulation environment, and results are reported.