Analysis of exit probability for a trajectory tracking robot in case of a rare event. Issue 4 (9th April 2022)

• Record Type: Journal Article
• Authors: Rana, Rohit; Gaur, Prerna; Agarwal, Vijyant; Parthasarathy, Harish
• Is Part Of: Robotica. Volume 40:Issue 4(2022)
• Page Start: 907
• Page End: 932
• Publication Date: 2022-04-09
• Subjects: nonlinear dynamics -- large deviation theory -- Poisson process -- Gaussian noise -- robot -- probability -- robotics; Robots -- Periodicals; 629.89205
• DOI: 10.1017/S0263574721000916
• Languages: English
• ISSNs: 0263-5747

Abstract

Abstract: In this paper, a novel statistical application of large deviation principle (LDP) to the robot trajectory tracking problem is presented. The exit probability of the trajectory from stability zone is evaluated, in the presence of small-amplitude Gaussian and Poisson noise. Afterward, the limit of the partition function for the average tracking error energy is derived by solving a fourth-order system of Euler–Lagrange equations. Stability and computational complexity of the proposed approach is investigated to show the superiority over the Lyapunov method. Finally, the proposed algorithm is validated by Monte Carlo simulations and on the commercially available Omni bundle TM robot.