Roll motion control with a fully active suspension system: High-Level Motion Control and Low-Level Actuator Implementation

Sammanfattning

During cornering, vehicle roll motion directly affects lateral load transfer, tire forcedistribution, and handling consistency. This thesis therefore investigates roll-gradientshaping for a fully active suspension system in simulated cornering maneuvers anddriver-in-the-loop tests, with emphasis on how body-level roll objectives can be realizedthrough actuator-level force generation.The proposed method combines a high-level body-motion controller with a low-levelhydraulic actuator controller. At the high level, lateral acceleration is estimatedfrom vehicle speed and steering input and used to generate a feedforward roll moment,while proportional-integral (PI) feedback is applied to roll, pitch, and heavestates. The resulting generalized force and moment targets are transformed intocorner suspension-force commands through a pseudoinverse force-allocation method.At the low level, the required damper forces are converted into continuously controlleddamper (CCD) current commands and motor-pump unit (MPU) flow-ratecommands using supplier-based lookup tables and mode-switching logic.The controller is implemented in MATLAB/Simulink and evaluated in both IPGCarMaker and the VI-CRT driver-in-the-loop real-time simulator. In CarMaker, theproposed controller tracks prescribed roll-gradient targets and produces suspensionforcedistributions that remain consistent with the intended vehicle behavior. Thelow-level controller also shows satisfactory force-tracking performance, with the maindeviations appearing near rapid force reversals. In the VI-CRT simulator, the controllerpreserves the expected roll characteristics over a wide lateral accelerationrange under real-time driving conditions, despite driver input variability and transientsteering effects.Overall, the results show that roll-motion control for a fully active suspension systemshould be evaluated as a connected body-level and actuator-level problem. Highlevelmotion objectives and low-level actuator behavior must therefore be consideredtogether when evaluating active suspension concepts.