Optimal control of vehicle random vibration with constrained suspension deflection

A major purpose of any vehicle suspension system is to isolate the body from roadway unevenness disturbances. On rough roads or at very high speeds on smoother roads, suspension deflection may become large enough to cause frequent bottoming, thus causing a severe degradation of isolation. Variable parameter suspensions are optimized to use the available suspension deflection to provide maximum isolation. Broadband stochastic roadway inputs at several intensity levels are applied to a quarter car model and the suspension parameters are optimized to find the best possible isolation under the (equality) constraint that the r.m.s. suspension deflection is a constant value in every case. From these results, improvements can be designed on the basis of measured suspension travel signals. Five types of suspension systems are investigated. These systems are called the fully active, the limited active, the optimal passive, the actively damped, and the variable damper systems. Comparisons are made among these systems in terms of r.m.s response, frequency domain predictions and eigen-frequency behavior as functions of disturbance intensity. The results show that such an adaptation philosophy would work well for moderate and high roadway intensities. For very smooth roads, forcing the suspension system to provide a specific r.m.s. travel does not produce a useful result, because at low disturbance intensities, the normal tire force variation becomes unnecessarily large as the system is forced to maintain constant r.m.s. suspension deflection. The fully active suspension system provides much better body isolation than the other types with or without equality constraint. Finally, the fully active system requires minimum suspension control force to maintain constant suspension travel compared with the other types of suspension considered. (C) 1996 Academic Press Limited