Skip to main navigation Skip to search Skip to main content

NRTIRL Based NN-RRT* Path Planner in Human-Robot Interaction Environment

  • Yao Wang
  • , Yuqi Kong
  • , Zhiyu Ding
  • , Wenzheng Chi*
  • , Lining Sun
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In human-robot interaction environment, it is of great significance for mobile robots to have the awareness of social rules, to realize the socialization and anthropomorphism of robot navigation behavior, and enhance the scene adaptation ability of socialized navigation. Learning from Demonstration (LfD) can obtain an optimized robot trajectory by learning the expert path. Inspired by the LfD method, we propose a new navigation method to learn navigation behaviors form demonstration paths of experts by Neural Network Rapidly-exploring Random Trees (NN-RRT*) planner in the human-robot interaction environment. First, we propose a new NN-RRT* planner to generate paths. Next, the features of demonstration paths and planned paths are extracted for Inverse Reinforcement Learning (IRL) process. The cost function of the path planner is updated. Finally, a new NN-RRT* that can adapt to the complex human-robot interaction environment is obtained. The experimental results show that comparing with the state-of-the-art methods, the path generated by the new navigation method has a higher degree of anthropomorphism and is suitable for navigation in a complex human-robot interaction environment.

Original languageEnglish
Title of host publicationSocial Robotics - 14th International Conference, ICSR 2022, Proceedings
EditorsFilippo Cavallo, Laura Fiorini, Alessandra Sorrentino, John-John Cabibihan, Hongsheng He, Xiaorui Liu, Yoshio Matsumoto, Shuzhi Sam Ge
PublisherSpringer Science and Business Media Deutschland GmbH
Pages496-508
Number of pages13
ISBN (Print)9783031246661
DOIs
StatePublished - 2022
Externally publishedYes
Event14th International Conference on Social Robotics, ICSR 2022 - Florence, Italy
Duration: 13 Dec 202216 Dec 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13817 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th International Conference on Social Robotics, ICSR 2022
Country/TerritoryItaly
CityFlorence
Period13/12/2216/12/22

Keywords

  • IRL
  • NN-RRT*
  • Robot navigation

Fingerprint

Dive into the research topics of 'NRTIRL Based NN-RRT* Path Planner in Human-Robot Interaction Environment'. Together they form a unique fingerprint.

Cite this