An iterative learning controller for nonholonomic mobile robots (original) (raw)
Related papers
Cyclic learning control of chained-form systems with application to car-like robots
1996
Abstract. An iterative learning controller for steering chained-form systems is presented. The learning algorithm relies on chained-form systems being linear under piecewise-constant inputs. The proposed control scheme requires the execution of a limited number of experiments in order to reach the desired state in finite time, with nice convergence and robustness properties with respect to modeling inaccuracies and disturbances.
Learning feedforward controller for a mobile robot vehicle
Control Engineering Practice, 1996
This paper describes the design and realisation of an on-line learning posetracking controller for a three-wheeled mobile robot vehicle. The controller consists of two components. The first is a constant-gain feedback component, designed on the basis of a second-order model. The second is a learning feedforward component, containing a singlelayer neural network, that generates a control contribution on the basis of the desired trajectory of the vehicle. The neural network uses B-spline basis functions, enabling a computationally fast implementation and fast learning. The resulting control system is able to correct for errors due to parameter mismatches and classes of structural errors in the model used for the controller design. After sufficient learning, an existing static gain controller designed on the basis of an extensive model has been outperformed in terms of tracking accuracy.
A neural network controller for a nonholonomic mobile robot with unknown robot parameters
Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292), 2002
In this paper, real-time fine motion control of a nonholonomic mobile robot is investigated, where both the robot dynamics and geometric parameters are completely unknown. A novel neural network controller combining both kinematic control and dynamic control is developed. The neural network assumes a single layer structure, by taking advantage of the robot regressor dynamics that express the highly nonlinear robot dynamics in a linear form in terms of the known and unknown robot parameters. The learning algorithm is computationally efficient. The system stability and the convergence of tracking errors to zero are rigorously proved using a Lyapunov stable theory. The real-time fine control of mobile robot is achieved through the on-line learning of the neural network. In addition, the developed controller is capable of learning the kinematic parameters on-line. The effectiveness and efficiency of the proposed controller is demonstrated by simulation studies.
Sliding Mode Control for Nonholonomic Mobile Robot
2000
A new control scheme is presented for nonholonomic mobile robots. The main idea of this paper is to consider the natural algebraic structure of the chained form system together with ideas from sliding mode theory while designing the control law. At first, the system model is converted into a single-input time-varying linear system by setting one input as a time-varying function. We design the sliding mode control law by using pole placement based on pseudo-linearized model. The point stabilization and path-tracking problem for chained form are studied based on these ideas. Simulations for both unicycle car and car like robot showed this control algorithm can make the mobile robot stabilized at desired configuration and following the desired trajectory with a high precision.
Sliding Mode Motion Control of Nonholonomic Mobile Robots
IEEE Control Systems, 1999
s nonholonomic mobile robots have constraints imposed on motions that are not integrable, i.e., the constraints cannot be written as time derivatives of some function of the generalized coordinates, advanced techniques are needed for the tracking control. In this paper a robust control law is proposed for trajectory tracking of nonholonomic wheeled mobile ro-One important topic that requires much attention (but has been studied little) is the problem of control of nonholonomic systems when there are model uncertainties. In most cases, the control problem is stated in terms of stabilizing a simple mathematical model, and the state variables are supposed to be known exactly at any time. However, taking into account several intrinsic characteristics of nonholonomic systems such as the actual bots. The state variables of the mobile robot are represented in polar coordinates, and the dynamic equation of the system is feedback-linearized by a computed-torque method. A novel sliding mode control law is derived for asymptotically stabilizing the mobile robot to a desired trajectory. It is shown that the proposed scheme is robust to bounded system disturbances. Simulation examples and experimental results are provided to show the effectiveness of the accurate tracking capability and the robust performance of the proposed controller. [ 1 11, proposed a sliding mode control that exploits a property named differential flatness of the kinematics of nonholonomic systems. In Guldner, et al. [ 121, a Lyapunov navigation function
Real-time motion control of a nonholonomic mobile robot with unknown dynamics
2001
In this paper, real-time fine motion control of a nonholonomic mobile robot is investigated, where the robot dynamics is completely unknown and is subject to significant uncertainties. A novel neural network based real-time controller is developed. The neural network assumes a single layer structure, by taking advantage of the robot regressor dynamics that express the highly nonlinear robot dynamics in a linear form in terms of the known and unknown robot dynamic parameters. The learning algorithm is computationally efficient. The system stability and the converge of tracking errors to zero are rigorously proved using a Lyapunov theory. The real-time fine control of mobile robot is achieved through the on-line learning of the neural network without any off-line learning procedures. In addition, the developed controller is capable of compensating sudden changes of robot dynamics or disturbance. The effectiveness and efficiency of the proposed controller is demonstrated by simulation studies.
Robust cascaded feedback linearizing control of nonholonomic mobile robot
In this paper, the problem of tracking control of nonholonomic mobile robot is investigated using feedback linearizing control. A cascaded control strategy has been designed to control the real mobile robot (kinematic model and robot dynamics), where the inner loop control for dynamics model is based on inverse dynamics and the outer loop control based on dynamic feedback linearizing control is carried out for kinematics model. The closed loop system is fully linearizable and described by a chain of integrators, and an exponentially stabilizing feedback for the desired trajectory can be designed for pole placement. The controller deals with unknown disturbance through a compensator carried out from the dynamics controller. The disturbance compensator contains an integral action, which eliminates the steady errors and enhances the robustness of the control scheme. Simulations are carried out for a nonholonomic mobile robot to verify the performance of the proposed control scheme.
2008 16th Mediterranean Conference on Control and Automation, 2008
In this paper, a trajectory tracking control for a nonholonomic mobile robot by the integration of a kinematic neural controller (KNC) and a torque neural controller (TNC) is proposed, where both the kinematic and dynamic models contains parametric and nonparametric uncertainties. The proposed neural controller (PNC) is constituted of the KNC and the TNC, and designed by use of a modeling technique of Gaussian radial basis function neural networks (RBFNNs). The KNC is applied to compensate the parametric uncertainties of the mobile robot kinematics. The TNC, based on the sliding mode theory, is constituted of a dynamic neural controller (DNC) and a robust neural compensator (RNC), and applied to compensate the mobile robot dynamics, significant uncertainties, bounded unknown disturbances, neural network modeling errors, influence of payload, and unknown kinematic parameters. To alleviate the problems met in practical implementation using classical sliding mode controllers and to eliminate the chattering phenomenon is used the RNC of the TNC, which is nonlinear and continuous, in lieu of the discontinuous part of the control signals p resent in classical forms. Also, the PNC neither requires the knowledge of the mobile robot kinematics and dynamics nor the time-consuming training process. Stability analysis and convergence of tracking errors to zero as well as the learning algorithms for weights are gua ranteed •with basis on Lyapunov method. Simulations results are provided to show the effectiveness of the proposed approach.
Learning optimal trajectories for non-holonomic systems
International Journal of Control, 2000
Many advanced robotic systems are subject to nonholonomic constraints, e.g., wheeled mobile robots, space manipulators and multifingered robot hands. Steering these mechanisms between configurations in the presence of perturbations is a difficult problem. In fact, the divide et impera strategy (first plan a trajectory, then track it by feedback) has a fundamental drawback in this case: due to the peculiar control properties of nonholonomic systems, smooth feedback cannot provide tracking of the whole trajectory. As a result, it would be necessary to give up either accuracy in the final positioning or predictability of the actual motion. We pursue here a different approach which does not rely on a separation between planning and control. Based on the learning control paradigm, a robust steering scheme is devised for systems which can be put in chained form, a canonical structure for nonholonomic systems. By overparameterizing the control law, other performance goals can be met, typically expressed as cost functions to be minimized along the trajectory. As a case study, we consider the generation of robust optimal trajectories for a car-like mobile robot, with criteria such as total length, maximum steering angle, distance from workspace obstacles, or error with respect to an off-line planned trajectory. * Corresponding Author
Neurocontrollers for Trajectory Tracking Problem of a Nonholonomic Mobile Robot
IFAC Proceedings Volumes, 2008
In this paper, a trajectory tracking control for a nonholonomic mobile robot by the integration of a kinematic controller and a torque controller is investigated. The proposed torque controllers (PTCs) are based on a Gaussian radial basis function neural network (RBFNN) modeling technique, which are used to compensate the mobile robot dynamics, significant uncertainties and disturbances. Also, the PTCs are not dependent of the robot dynamics neither requires the off-line training process. The stability analysis and the convergence of tracking errors to zero, as well as the learning algorithms (for weights, centers, and widths) are guaranteed with basis on Lyapunov's theory. In addition, the simulation results shows the efficiency of the PTCs.