Multilayer Perceptron Adaptive Dynamic Control for Trajectory Tracking of Mobile Robots

Abstract

This paper presents a novel functional-adaptive dynamic controller for trajectory tracking of nonholonomic wheeled mobile robots. The controller is developed in discrete-time and employs a multilayer perceptron neural network for the estimation of the robot's nonlinear dynamic functions, which are assumed to be completely unknown. On-line weight tuning is achieved by employing the extended Kalman filter algorithm, based on a specifically formulated stochastic inverse dynamic identification model of the mobile base. A discrete-time dynamic control law employing the estimated functions is proposed and cascaded with a trajectory tracking kinematic controller. The performance of the complete system is analysed and compared by realistic simulations.

Dual Adaptive Dynamic Control of Mobile Robots Using Neural Networks

“…2) Sigmoidal MLP Dual Adaptive Scheme: Another type of neural network, commonly employed in control applications, is the sigmoidal MLP ANN [7], [14], [16], [45]. Unfortunately, MLP networks do not preserve the desirable property of linearity in the unknown network parameters exhibited by radial basis function networks.…”

“…Parametric adaptive control and robust sliding mode control, have also been proposed to mitigate the problem of unknown or uncertain mobile robot parameters [8], [10]. Another approach is that of online functional-adaptive control, where the uncertainty is not restricted to parametric terms, but covers the dynamic functions themselves [7], [11], [14] [17]. We consider this approach to be more general and superior in handling higher degrees of uncertainty and unmodelled dynamics.…”

“…Adaptive controllers which have hitherto been proposed for the control of mobile robots, rely on the heuristic certainty equivalence (HCE) assumption [7], [11], [14] [17]. This means that the estimated functions are used by the controller as if they were the true ones, thereby ignoring completely their uncertainty.…”

Multilayer perceptron dual adaptive control for mobile robots

“…Definition 3.6: ∇ h k denotes the Jacobian matrix of h (x k−1 , τ k−1 , z * k ) with respect to z * k evaluated atẑ k . By (9), (10) and (12) it can be shown that:…”

Enhanced SLAM for a mobile robot using extended Kalman Filter and neural networks

“…[9][10][11][12][13] Houshangi and Azizi 7 integrated the information from odometry and gyroscope using UKF. To improve the performance of odometry, a fiber optic gyroscope is used to give the orientation information that is more reliable.…”