Multi-Robot Collaborative Autonomous Exploration: Efficiently Mapping an Unknown Area
Typ
Examensarbete för masterexamen
Master's Thesis
Master's Thesis
Program
Systems, control and mechatronics (MPSYS), MSc
Complex adaptive systems (MPCAS), MSc
Complex adaptive systems (MPCAS), MSc
Publicerad
2023
Författare
Wallström, Anders
Brask, Edward
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
Autonomous exploration carried out by multiple mobile robots offers significant advantages in efficiently mapping areas. With advancements in communication technologies, collaborative mapping using multiple robots has become a feasible approach. In line with this, Ericsson has supported this thesis project which researches and develops a system using these technologies. This master’s thesis develops a scalable and collaborative multi-robot system for autonomous exploration. It combines the research domains of multi-robot Simultaneous Localization And Mapping (SLAM), autonomous exploration, and map merging to efficiently create a map of an unknown environment. The system was implemented in the Robot Operating System (ROS) and used node-based exploration strategies combined with computer vision techniques to perform collaborative exploration. This thesis serves as a proof of concept and can provide a framework for future research and development in this field. One potential application for collaborative exploration is search and rescue operations. The collaborative aspects of the system have been effectively implemented and function well. Through extensive hardware testing in three distinct environments, the results demonstrate a significant in- crease in efficiency and lower exploration time with the collaborative system compared to a single robot exploring autonomously. However, there are accumulating errors within the system, limiting the SLAM accuracy compared to a single robot. Furthermore, the system’s performance is highly dependent on the environment and the initial positions of the robots, making it challenging to predict absolute performance and determine optimal parameters in unknown environments. Future work involves enhancing collision avoidance capabilities, dynamically adjusting exploration parameters based on the characteristics of the unknown environment, and transitioning from in- dependent SLAM algorithms to an integrated multi-robot SLAM system to improve collaborative accuracy.
Beskrivning
Ämne/nyckelord
Multi-robotsystem , Collaborativeexploration , Autonomousexploration , SLAM , Mapmerging , Scalableframework, , Autonomousrobots