
Towards Development of Performance Metrics for Benchmarking SLAM Algorithms
Author(s) -
Mudit Bhargava,
Rushad Mehta,
Chandan Das Adhikari,
K Sivanathan
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1964/6/062115
Subject(s) - simultaneous localization and mapping , benchmarking , computer science , mobile robot , artificial intelligence , robot , robotics , motion planning , computer vision , marketing , business
The true autonomy of mobile robots cannot be achieved without Simultaneous Localization and Mapping (SLAM). With this capability, mobile robot could concurrently build a map of the environment and locate itself with respect to the map. Although there are several variants of SLAM algorithms contributed by researchers so far, only a very few works were aimed at comparing their performances with appropriate metrics and providing detailed directions and insights to the user on selection criteria and indicative use cases. In this work, we presented a comparative study of three popular SLAM algorithms and provide some significant quantitative performance measures of the same by using our novel | R | and | S | performance metrics as well as conventional metrics. The comparative study was carried out in ROS (Robot Operating System) using Turtlebot3 robot model on three SLAM packages viz G-mapping, Karto SLAM, and Frontier Exploration SLAM. Furthermore, the results show that the proposed metrics are very efficient and compact in comparing and quantifying the performance of SLAM algorithms.