R-OAM: A reliable method for odometry accuracy measurement

This paper presents R-OAM, a reliable method for odometry accuracy measurement. Compared to existing algorithms, R-OAM demonstrates enhanced stability and reliability in estimating the accuracy of odometry in both translation and rotation. R-OAM aligns trajectories by establishing a constraint equation that considers both translation and rotation components, effectively eliminating alignment errors caused by initial trajectory differences and providing a more precise representation of drift. Additionally, R-OAM incorporates a Euler angle singular value filtering mechanism to enhance the reliability of rotation evaluation. To facilitate the evaluation process, an evaluation tool has been developed based on the proposed method, automating the generation of plotting scripts. Experimental validation demonstrates that R-OAM effectively overcomes difficulties related to attitude evaluation and significantly improves trajectory alignment. Consequently, the integration of R-OAM with its evaluation tool presents an accurate, efficient, and user-friendly approach for measuring odometry accuracy.