Ricardo López García

Neural Control of the Movements of a Wheelchair

This paper studies the problem of controlling the movements of a handicapped persons motorized wheelchair from a practical point of view. The control system implemented has been divided into two levels: the “low level”, consisting of an electronic system which directly controls the drivers of the chairs motors, with a classic PID (proportional-integral-derivative) control loop. The aim of this level is to ensure that the speeds of each one of the wheels is similar to the input speed of these control boards. The second control level (“high level”), implemented by means of neural techniques, ensures that the linear and angular speeds of the wheelchair are those indicated by a trajectory generator. A new recurrent model is used as the neural network, for which the stability conditions of the complete control system are obtained and various practical tests are carried out, which show the correct performance of the actual system implemented.

Identification and control of a wheelchair using recurrent neural networks

Abstract This work involves the control of a wheelchair using a new model of radial base function (RBF) recurrent neural networks. The proposed architecture is made up of two blocks, each with one neural network: one to identify the physical system (plant)—the identifier, and another for control—the controller. The identifier, running in parallel with the plant, is designed to obtain the system’s Jacobian, which is used to adjust the weights of the controller. The stability conditions are obtained for the correct functioning of the system, and several tests are described in which the movements of a wheelchair are governed, thus confirming the correct functioning of the control architecture used.

Hardware implementation of a new neurocontrol wheelchair-guidance system

Abstract This paper describes the implementation of a neural network control system for guiding a wheelchair, using an architecture based on a digital signal processor (DSP). We use a recurrent radial basis neural network as a system controller and a Kalman filter as identifier, which acts as propagator of the control errors to the neurocontroller. The hub of the algorithm architecture is a DSP of the company Texas Instruments (TMS320C31). The board has complete autonomy of action and is specifically designed for executing control algorithms in real time. Various practical tests have been carried out to check the correct functioning of the system when governing the output of the wheelchair.

Advanced and Intelligent Control Techniques Applied to the Drive Control and Path Tracking Systems on a Robotic Wheelchair

This paper presents the theoretical support and experimental results of the application of advanced and intelligent control techniques to the drive control and trajectory tracking systems on a robotic wheelchair. The adaptive optimal control of the differential drive helps to improve the automatic guidance systems safety and comfort taking into consideration operating conditions such as load and distribution changes or motion actuator limitations. Furthermore, the incorporation of an optimal controller to minimize location errors and a fuzzy controller to adapt the linear velocity to the characteristics of the trajectory, provide the vehicle with a high degree of intelligence and autonomy, even when faced with obstacles. The global control solution implemented increases the features of the wheelchair for handicapped people, especially for those with a high degree of disability.

Identification of Systems Using Radial Basis Networks Feedbacked with FIR Filters

A new model of a radial basis neural network is presented in this article which is fedbacked with a FIR filter. Using various neurons of this type, it is possible to construct a recurrent neural network, where the coefficients of each filter and the synaptic connections are adjusted to minimize an error function. The simulations carried out show the validity of this method for identifying systems with memory.

Ego-motion computing for vehicle velocity estimation

In this paper, we present a method for computing velocity using a single camera onboard a road vehicle, i.e. an automobile. The use of computer vision provides a reliable method to measure vehicle velocity based on egomotion computation. By doing so, cumulative errors inherent to odometry-based systems can be reduced to some extent. Road lane markings are the basic features used by the algorithm. They are detected in the image plane and grouped in couples in order to provide geometrically constrained vectors that make viable the computation of vehicle motion in a sequence of images. The applications of this method can be mainly found in the domains of Robotics and Intelligent Vehicles.

Using a New Model of Recurrent Neural Network for Control

This paper shows the results obtained in controlling a mobile robot by means of local recurrent neural networks based on a radial basis function (RBF) type architecture. The model used has a Finite Impulse Response (FIR) filter feeding back each neurons output to its own input, while using another FIR filter as a synaptic connection. The network parameters (coefficients of both filters) are adjusted by means of the gradient descent technique, thus obtaining the stability conditions of the process. As a practical application the system has been successfully used for controlling a wheelchair, using an architecture made up by a neurocontroller and a neuroidentifier. The role of the latter, connected up in parallel with the wheelchair, is to propagate the control error to the neurocontroller, thus cutting down the control error in each working cycle.

Modelling and Simulation of the Kinematic and Dynamic Behavior of a Fork-Lift-Truck

Abstract This paper presents the kinematic and dynamic model of a fork-lift-truck with one drive-steering wheel and two fixed and passive wheels. The simulation of the model vehicle allows the analysis of its behavior in the face of variations in traction and direction including effects such as changes in the load, friction due to the wheel-surface contact, etc. Furthemore, as algorithm integrated in the control and guidance structure of the fork-lift-truck, the model obtained allows the link between the path generator-tracker system and the direct control of the motors associated to the active wheel.

Identification and Adaptive Control of Systems Using Recurrent Neural Networks

Abstract This work involves the adaptive control of nonlinear systems using a new model of radial base function (RBF) recurrent neural networks (Ciocoiu, 1996). The proposed architecture is made up of two blocks, each with two neural networks, one to identify the physical system (plant) - identifier, and another for control - controller. The identifier, running in parallel with the plant, is designed to obtain the system’s Jacobian, which is used to adjust the weights of the controller. Simulations carried out on various types of plant prove that the method works well.

Navigation System for a Mobile Robot by Means of an Infrared System

Abstract This paper describes how an infrared system could improve a mobile robot following a path using an odometric system. The infrared system takes out the accumulative errors given by the odometrics, and the tracking errors are minimized using a Kaiman filter.