An Improved SOM-Based Method for Multi-Robot Task Assignment and Cooperative Search in Unknown Dynamic Environments

The methods of task assignment and path planning have been reported by many researchers, but they are mainly focused on environments with prior information. In unknown dynamic environments, in which the real-time acquisition of the location information of obstacles is required, an integrated multi-r...

Full description

Saved in:
Bibliographic Details
Published inEnergies (Basel) Vol. 13; no. 12; p. 3296
Main Authors Tang, Hongwei, Lin, Anping, Sun, Wei, Shi, Shuqi
Format Journal Article
LanguageEnglish
Published Basel MDPI AG 01.06.2020
Subjects
Algorithms
Computer simulation
dynamic window approach
Hovering
Locking
Methods
multi-robot
Multiple robots
Neural networks
Neurons
Obstacle avoidance
Path planning
Planning
Real time
Researchers
Robots
Searching
Self organizing maps
self-organizing map
Sensors
task assignment and cooperative search
unknown dynamic environments
Velocity
Wireless networks
Online AccessGet full text

Cover

More Information
Summary:The methods of task assignment and path planning have been reported by many researchers, but they are mainly focused on environments with prior information. In unknown dynamic environments, in which the real-time acquisition of the location information of obstacles is required, an integrated multi-robot dynamic task assignment and cooperative search method is proposed by combining an improved self-organizing map (SOM) neural network and the adaptive dynamic window approach (DWA). To avoid the robot oscillation and hovering issue that occurs with the SOM-based algorithm, an SOM neural network with a locking mechanism is developed to better realize task assignment. Then, in order to solve the obstacle avoidance problem and the speed jump problem, the weights of the winner of the SOM are updated by using an adaptive DWA. In addition, the proposed method can search dynamic multi-target in unknown dynamic environment, it can reassign tasks and re-plan searching paths in real time when the location of the targets and obstacle changes. The simulation results and comparative testing demonstrate the effectiveness and efficiency of the proposed method.
ISSN:1996-1073
1996-1073
DOI:10.3390/en13123296