Search Results

This paper focuses on the design of a self-optimizing nonlinear controller for a “simplified” pneumatic brake system in the continuous time domain. The specific controller under investigation periodically excites the brake system in the direction towards the maximum tire forces without apriori knowledge on the initial direction of motion. The nonlinearity and complexity of pneumatic systems (as opposed to hydraulics) introduce a higher level challenge. The developed controller employs multiple observers to estimate tire forces in a highly unpredictable environment with bounded parameter uncertainties. The controller explicit inputs are the wheel speeds and chamber pressures. A longitudinal accelerometer is also recommended.

The primary objective of this paper is to develop a model that accurately represents the dynamics of air flowing through the components of a pneumatic system configuration, which is common in many heavy duty vehicle applications, that eventually translates into braking force. This objective is met using the dynamic compressible airflow equations, which describe flow through an orifice. These equations are coordinated to describe the directional motion of dynamic airflow as commanded by the driver at the foot-pedal and as modified downstream by a modulator to facilitate ABS activity. The solenoid actuated relay valve also includes the motion dynamics of a piston in the existence of hysteresis and coulomb friction type built-in non-smooth nonlinearities. The adoption of an isentropic process, as opposed to the more general case of polytropic behavior, is experimentally determined to suffice for accuracy while yielding significant mathematical convenience.

This paper discusses the needs for utilizing pure simulation based studies in the algorithm development process for cost reduction in today's competitive business environment. Continuous improvement to any algorithm is essential for sustained existence and any modification to the logic needs to be verified to work for every scenario from feasibility and robustness viewpoints. A Pareto analysis performed on the development cycle of an ABS algorithm identifies these high cycles clearly. In this paper, it is proposed that significant cycle time reduction can be achieved with the aid of pure desktop simulations during the development phases. Additionally, the feasibility of using such simulations is demonstrated through the ABS algorithm simulation model that is developed at AlliedSignal Truck Brake Systems.

This paper is concerned with Advanced Transportation Systems, in particular, the design of controllers for Fully-Automated Vehicle Operation. We specifically consider the design and implementation of a lateral controller for a cooperative vehicle system being developed at The Ohio State University. The objective of the lateral controller is to steer the vehicle to follow a retroreflector stripe placed on the roadway pavement using radar sensors. The structure and the parameters of the controller are determined during simulations and analytic studies. The controller models are then downloaded into two high-speed computer systems which are interconnected to simulate the operation of the closed loop system in real time and provide a “hardware-in-the-loop” environment. Finally, the computer containing the controller dynamics is installed in the vehicle and field experiments are conducted.