Social robot navigation must consider not only task constraints, such as the minimum path length, but also social conventions, such as satisfying the social acceptability of the path. This paper presents a new strategy for social robot navigation based on 2D Gauss-Gumbel spatial density function to consider the human state (position, direction and motion) and social interaction information related to robots, which model the personal space and social interaction space respectively. The personal space and social interaction space constitute a Dynamic Social Space (DSS). The DSS based human comfort and safety navigation can estimate the approaching pose of a robot for a person or a group of people, so the robot can ensure not only people's safety but also comfort when approaching a person or group of people in social situations. We evaluate the developed model through simulation and real-world experiments using the newly proposed social individual comfort index and social group comfort index.
Social Robot Navigation Based on a 2D Gauss-Gumbel Spatial Density Model in Human-Populated Environments
Lect. Notes Electrical Eng.
The International Conference on Image, Vision and Intelligent Systems (ICIVIS 2021) ; Chapter : 99 ; 1133-1143
2022-03-03
11 pages
Article/Chapter (Book)
Electronic Resource
English
Human-like motion planning in populated environments
TIBKAT | 2019
|Robot navigation in human environments
TIBKAT | 2019
|COACHES Cooperative Autonomous Robots in Complex and Human Populated Environments
BASE | 2015
|BASE | 2013
|BASE | 2013
|