MAES 2.0: A ROS Compatible Simulation Tool for Multi Robot Exploration and Coverage
Authors
Andreasen, Malte Zoëga ; Jensen, Magnus Kirkegaard ; Holler, Philip Irming
Term
4. term
Education
Publication year
2022
Pages
22
Abstract
Multi Agent Exploration Simulator (MAES) is an open-source, physics-based simulator built in Unity that lets multiple robots move and sense in a virtual world using discrete time steps. We present MAES 2.0 as an extension of MAES 1.0, adding heatmap visualization (to reveal patterns), configurable signal degradation, support for user-provided maps, and an expanded debugging GUI. MAES 2.0 also offers a ROS (Robot Operating System) interface, making it easier to transfer algorithms from simulation to real robots. The codebase includes unit and system tests to ensure intended behavior. On modest hardware, performance tests showed that MAES 2.0 can simulate up to 5 robots in ROSMode (with ROS integration) and up to 120 robots in UnityMode. A usability test found that robotics researchers and developers could quickly install, set up, and use MAES to implement simple robot logic.
Multi Agent Exploration Simulator (MAES) er en open source, fysikbaseret simulator i Unity, der lader flere robotter bevæge sig og sanse i en virtuel verden i diskrete trin. Vi præsenterer MAES 2.0 som en udvidelse af MAES 1.0, der tilføjer heatmap-visualisering (for at afsløre mønstre), brugerdefineret signaldæmpning, support for brugerleverede kort samt en udvidet grafisk brugerflade til fejlsøgning. MAES 2.0 kan desuden forbindes til ROS (Robot Operating System), hvilket gør det lettere at overføre udviklede algoritmer fra simulation til fysiske robotter. MAES 2.0 indeholder både enhedstests og systemtests for at sikre den ønskede adfærd. Ydeevnetests på beskeden hardware viste, at MAES 2.0 kan simulere op til 5 robotter i ROSMode (med ROS-integration) og op til 120 robotter i UnityMode. En brugervenlighedstest viste, at målgruppen af robotikforskere og -udviklere hurtigt kunne installere, sætte op og bruge MAES til at implementere simpel robotlogik.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords
robots ; robot ; swarm ; ros ; ros2 ; simulation ; maes ; distributed systems ; area exploration ; multi agent ; unity ; online coverage ; continuous space