Interpretable Fuzzy Logic Control for Multirobot Coordination in a Cluttered Environment

Mobile robot navigation is an essential problem in robotics. We propose a method for constructing and training fuzzy logic controllers (FLCs) to coordinate a robotic team performing collision-free navigation and arriving simultaneously at a target location in an unknown environment. Our FLCs are org...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on fuzzy systems 2021-12, Vol.29 (12), p.3676-3685
Main Authors: Chang, Yu-Cheng, Shi, Ye, Dostovalova, Anna, Cao, Zehong, Kim, Jijoong, Gibbons, Daniel, Lin, Chin-Teng
Format: Article
Language:English
Subjects:
Collision avoidance
Fuzzy control
Fuzzy logic
Fuzzy logic controller (FLC)
Fuzzy sets
Genetic algorithms
Multi-robot systems
multiobjective hybrid genetic algorithm-particle swarm optimization (GA-PSO)
Multiple robots
multirobot navigation
Navigation
Obstacle avoidance
Particle swarm optimization
Robotics
simultaneous arrival
Training
Unknown environments
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Mobile robot navigation is an essential problem in robotics. We propose a method for constructing and training fuzzy logic controllers (FLCs) to coordinate a robotic team performing collision-free navigation and arriving simultaneously at a target location in an unknown environment. Our FLCs are organized in a multilayered architecture to reduce the number of tunable parameters and improve the scalability of the solution. In addition, in contrast to simple traditional switching mechanisms between target seeking and obstacle avoidance, we develop a novel rule-embedded FLC to improve the navigation performance. Moreover, we design a grouping and merging mechanism to obtain transparent fuzzy sets and integrate this mechanism into the training process for all FLCs, thus increasing the interpretability of the fuzzy models. We train the proposed FLCs using a novel multiobjective hybrid approach combining a genetic algorithm and particle swarm optimization. Simulation results demonstrate the effectiveness of our algorithms in reliably solving the proposed navigation problem.
ISSN:1063-6706
1941-0034
DOI:10.1109/TFUZZ.2021.3111446