Approximating Kinematics for Tracked Mobile Robots

Abstract

In this paper we propose a kinematic approach for tracked mobile robots in order to improve motion control and pose estimation. Complex dynamics due to slippage and track-soil interactions make it difficult to predict the exact motion of the vehicle on the basis of track velocities. Nevertheless, real-time computations for autonomous navigation require an effective kinematics approximation without introducing dynamics in the loop. The proposed solution is based on the fact that the instantaneous centers of rotation (ICRs) of treads on the motion plane with respect to the vehicle are dynamics-dependent, but they lie within a bounded area. Thus, optimizing constant ICR positions for a particular terrain results in an approximate kinematic model for tracked mobile robots. Two different approaches are presented for off-line estimation of kinematic parameters: (i) simulation of the stationary response of the dynamic model for the whole velocity range of the vehicle; (ii) introduction of an experimental setup so that a genetic algorithm can produce the model from actual sensor readings. These methods have been evaluated for on-line odometric computations and low-level motion control with the Auriga-α mobile robot on a hard-surface flat soil at moderate speeds.

Adaptive Trajectory Tracking Control of Skid-Steered Mobile Robots

“…Song et al [2] use the tracked vehicle models discussed in [3]. In [4], localization of a tracked vehicle based on kinematic models is presented. For skidsteered modeling of tracked vehicles, readers can refer to [5]- [7] for details.…”

IMU-based localization and slip estimation for skid-steered mobile robots

“…We denote the instantaneous center of rotation (ICR) of the left-side wheel contact points, right-side wheel contact points, and the robot body as ICR l , ICR r , and ICR G , respectively. It is known that ICR l , ICR r and ICR G lie on a line parallel to the x-axis [8], [11], [15]. We denote the x-y coordinates for ICR l , ICR r , and ICR G as (x l , y l ), (x r , y r ), and (x G , y G ), respectively.…”

“…We denote the x-y coordinates for ICR l , ICR r , and ICR G as (x l , y l ), (x r , y r ), and (x G , y G ), respectively. We find that the y-coordinate S of the ICRs satisfies the following constraints [8], [15].…”

“…The value of S is considered as a function of variable γ. If ω L = ω R , γ = 0 and S is a finite value due to the non-zero lateral velocity of mass center G [8]. If ω L = −ω R , γ → ∞ and S = 0 since the robot is rotating about G without any translational motion.…”

“…In [6], [7], reduction of the odometry of the skid-steered robots are discussed for dead-reckoning applications using encoder and motor current measurement information. In [8], localization of a tracked vehicle is presented based on kinematic models provided in [9]- [11]. In [12], an IMU-based wheel slip detection scheme is designed for a unicycle-like mobile robot using an extended Kalman filter (EKF) with or without GPS measurements.…”