Kinematic modelling of tracked vehicles by experimental identification

Abstract

Abslr4cf -The paper proposes a ldnematic approach for tracked vehicles in order to improve motion control and pose estimation. Complex dynamics due to slippage and soil shearing make it difficult to predict the exact motion of the vehicle from the velocity of the two backs. Nevertheless, reliable geomeMc approxlmations are necessary to perform onboard real-time computations for autonomous navigation. The presented solution is based on the ldnematic similarities between tracked vehicles and wheeled differential drive vehicles. Particularly, the approximate position of wheel contact paints for am equivalent vehicle c m be opiimlzed for a particular terrain at moderate speeds. This is achieved off-line by feeding a genetic algorithm with raw trajectoIy data and reliable localization estimations based on external senson. The method bas been successfully tested for online odometric computations m d low-level motion control with the Aurigaa mobile robot Moreover, tbe identilied parameters are similar to those obtained from tbe simulated stationary response of a complex dynamic model of tbh vehlcle.

Design and implementation of a new teleoperation control mode for differential drive UGVs

“…First, some additions to the scheme presented in Section 3 must be made to account for the modeling errors induced by the slippage of the tracks, [16][17][18]. These additions are twofold: The kinematic model (3) is slightly extended, and a gyro feeding a Kalman filter is added to measure errors in UGV rotation.…”

“…We will apply the approach proposed in [18], which boils down to a slight modification of Equation (3). Instead of measuring the parameters L and d of Figure 5 we now introduce parameters L 1 , L 2 and d, as in Figure 9, and identify those parameters from experiments, yielding the results in Table 1 However, running the FLC and commanding a straight line camera translation the resulting trajectory overshoots and oscillates heavily when running at high speeds.…”

Development of a variable parallelogram tracked mobile robot

“…The traditional tracked mobile robot can be classified into tracked robot with fixed structure (single pair of track) and variable structure [4,5] . According to the number of variable arms, tracked robot can be divided into four-track two-arm robot and six-track four-arm robot.…”

Effect of Fast Motion on Range Images Acquired by Lidar Scanners for Automotive Applications

“…Moreover, we use a simple motion model since a car's state does not change significantly during one scan frame. In contrast to Martínez et al [12] the estimate of the car's pose requires only a single scan frame.…”

“…However, in contrast to Ono et al he even considered relative motion during one scan by synchronizing every measured point to the estimated trajectory [best estimated trajectory (BET)]. Finally, there is extensive work on the estimation of poses and motion parameters of robots using a sequence of lidar frames, e.g., [12].…”