Search Results

This paper describes a multi-degree-of-freedom model of a three-wheeled car implemented in Matlab®. The purpose was to investigate the dynamics of the car (assumed to be rigid on its suspension) during cornering. While the problems associated with three-wheeled cars are well known, much of the guidance in the literature and off-the-shelf software assumes a conventional four-wheeled car. Consequently, the authors were approached with a battery-electric concept car which was thought to offer better performance than the existing variants because the use of hub motors lowered the center of gravity and, hence, reduced rollover coefficient. However, simulation of the vehicle model in cornering shows that the concept is still prone to instability. Indeed, it suffers greater roll velocities than a comparable three-wheeled car with an internal combustion engine (ICE) because the ratio of sprung to unsprung mass is significantly altered.

The use of electric motors to independently control the torque of two or four wheels of a vehicle has the potential to significantly improve safety and handling. One virtue of electric motors is that their output torque can be accurately estimated. Using this known output torque, longitudinal tire force and coefficient of friction can be estimated via a controller output observer. This observer works by constructing a model of wheel dynamics, with longitudinal tire force as an unknown input quantity. A known wheel torque is input to the physical and modeled system and the resulting measured and predicted wheel speeds are compared. The error between the measured and predicted wheel speed is driven towards zero by a robust feedback controller. This controller modulates an estimate of longitudinal tire force used as an input by the wheel dynamics model. The resulting estimate of longitudinal tire force quickly converges towards the actual value with minimal computational expense.

This paper presents a driver assistance system designed to minimize the effect of driver reaction time on lane and speed maintenance operations. Nearly-instantaneous correcting actions are provided through a hierarchical arrangement of behaviors, by avoiding the time lag associated with deliberative or planning steps found in many control algorithms. Concepts originating in the field of robotics, including artificial potential fields and behavior-based systems, are interpreted for application to automotive control, where vehicle dynamics places considerable practical constraints on implementation. Ideas found in the study of emergent behavior in nature provide continuous, non-stepwise control signals, suitable for additive corrective inputs at highway velocities. This approach is effective for a substantial subset of road automobiles operating over a variety of speeds.

Vehicle stability augmentation has been refined over many years, and currently there are commercial systems that control right/left braking and throttle to create vehicles that remain controlled when road conditions are very poor. These systems typically use yaw rate and lateral acceleration in their control philosophy. The tire/road friction coefficient, μ, has a significant role in vehicle longitudinal and lateral control, and there has been associated efforts to measure or estimate the road surface condition to provide additional information for the stability augmentation system. In this paper, a differential braking control strategy using yaw rate feedback, coupled with μ feedforward is introduced for a vehicle cornering on different μ roads. A nonlinear 4-wheel car model is developed. A desired yaw rate is calculated from the reference model based on the driver steering input.

In the simulation of the dynamic response of a vehicle, the accuracy of the predictions strongly depends on the tire properties. Since the physics of tire force generation is highly nonlinear and complex, semi-empirical models are used, which are mathematically curve fitted to experimental data. Although this approach yields realistic tire behavior, it requires many experimental coefficients. Even though tire forces generated by a real tire are nonlinear, there is a linear region where the slip and slip angle are low. Most normal driving is done in this region. This paper will present a new analytical tire model capable of simulating pure cornering, pure braking, and combined braking/cornering in this region. The dynamic properties of the tire are analytically derived as functions of the slip, slip angle, normal force, and road friction coefficient.

A laser sensor has been developed which can measure, at high bandwidth, the distance between itself and some target. This sensor is proposed to be mounted at a fixed orientation to a vehicle equipped with automatic steering capability. It is assumed that the only available feedback variable is the distance from the sensor to a roadside target. Our objective was to design a controller for trajectory tracking and to determine the best orientation angle of the sensor. Analytical and simulation results indicate that a high quality of control with small trajectory error and acceptable lateral acceleration can be achieved with a sensor orientation between 20° and 45° with respect to the longitudinal vehicle axis.

It is commonly accepted that the principal functions of an automobile suspension are to control low frequency rigid body motions, provide comfort to passengers, and to reduce tire normal force variation so that predictable handling is maintained. A good argument for reducing normal force variation is that in the extreme, if a tire is off the ground, it for certain cannot generate any lateral forces, and thus compromises lateral dynamics. The direct relationship between road holding and dynamic tire normal force variation is quantified sparsely in the literature. In this paper a relatively simple model is proposed which exposes how normal force variation at the front and rear directly affects the vehicle yaw rate and lateral acceleration. It is shown that normal force variation at the front has potentially the same effect on lateral dynamics as does the steering input.

Semi-active suspension control deals principally with high bandwidth modulation of passively generated damper forces. When a properly designed semi-active damper is used in an otherwise conventional suspension there is much evidence that a superior overall system results. Large off-road vehicles, such as military transport vehicles, traveling at high speeds over rough terrain, possess suspension control requirements which are different from a road going vehicle. This paper develops these control requirements.

Large vehicles such as buses and trucks suffer from a dynamic resonance problem due to flexure of the vehicle frame. This type of vibration is called “beaming”. Unlike smaller, (relatively) stiffer automobiles, this flexural vibration occurs in these large vehicles at a frequency very close to the wheel-hop frequency and body motion frequencies, and thus is excited easily by typical roadway inputs. The effect of beaming on ride quality differs from one vehicle type to another due to the mounting of the passenger compartment to the frame. A model that provides insight into the beaming problem is formulated using bond graphs. Bond graphs allow coupling of lumped dynamic effects with the distributed dynamics of the frame, and they permit virtually automated computer coding so that different vehicle types can be simulated using basically the same model. The model utility is demonstrated for some simple automatic control strategies aimed at controlling beaming.

The effect of exhaust system configuration upon performance of two-stroke engines is explored. Computer predictions of gas dynamic behavior in the exhaust pipes are compared to experimental results of real pipe tests. Predicted pressure, velocity and temperature histories at key points in the exhaust system explain the relative power differences for the test pipes. Predicted volumetric flow rates show the effectiveness of exhausting gases at larger pipe cross sectional areas in reducing noise output. The application of the computer predictions to a Yamaha RS 100 offers an explanation for the experimentally measured loss of power at mid-range engine speeds encountered using one of the two test pipes. A relative comparison of the two systems shows that the differing pipe geometries cause predicted pressure and velocity histories to change. These changes are correlated to actual power curves.

The relation between the power and the pressure-time history at the exhaust port of a high performance two-stroke engine is explored. Through the use of a variable length expansion chamber with five different tailpipes, the effect on power of stuffing pulse arrival time, magnitude and duration was found. The results from this set of tests are compared with those of two side exit expansion chambers and an inverted tailpipe expansion chamber. Tests were also made to determine how mid-range torque and high speed torque effect of the performance of the motorcycle. An optimum configuration (pipe length and tail pipe length) was found for the 360 Yamaha tested.

The important internal dynamic interactions which ultimately determine the performance of a two-stroke internal combustion engine are identified and then modeled using the bond graph modeling technique. Each of the important dynamic characteristics are developed independently, presenting both the bond graph and mathematical formulation of each of the assumed subsystems. These subsystems are assembled into the final overall system bond graph and governing state space equations are derived. The result of this development is a reasonably straightforward procedure for assembling virtually any size two-stroke ICE with virtually any internal geometry into a cogent, simulatable mathematical formulation. The procedure outlined above is demonstrated for the Yamaha 360 MX motorcycle engine. In addition to presenting results for various internal dynamic interactions (pressure, flow, temperature, etc.), the model is shown to yield excellent predictions for power and torque as a function of RPM.