Articulated vehicle stability control using brake-based torque vectoring

Abstract

Statistics show that unstable articulated vehicles pose a serious threat to the occupants driving them as well as the occupants of the vehicles around them. An articulated vehicle typically experiences three types of instability: snaking, jack-knifing and rollover. An articulated vehicle subjected to any of these instabilities can result in major accidents. It is also known that many individuals are unaware of how to properly tow or pack a loaded articulated vehicle. These individuals are, therefore, at a high risk of causing the vehicle system to become unstable. It can hence be confidently said that a method in which an articulated vehicle can stabilise itself is a worthy research question. The method that is implemented in this study is to create a control system, using Nonlinear Model Predictive Control (NMPC), that has the capability of stabilising an articulated vehicle by applying torque vectoring to the trailer. In order for this control system to be applied, a nonlinear articulated vehicle MSC ADAMS model was constructed. The NMPC controller works by using a nonlinear explicit model to predict the future states of the vehicle and then finding the optimal left and right braking forces of the trailer by minimising the cost function using least squares minimisation. The cost function includes the towing vehicle yaw rate, trailer yaw rate and hitch angle and is minimised by minimising the error between the desired vehicle states and the actual states. It was found that the NMPC is capable of not only preventing instability but also causes the vehicle system to behave as if the trailer is unloaded. This conclusion means that this type of control system can be used on all types of articulated vehicles and shall ensure the safety of not only the vehicle occupants but other road users as well.

Unfortunately, due to the impact of the 2020 COVID-19 pandemic, the experimental validation of the model had to be delayed significantly. It is for this reason that the experimental validation for the controller could not be done.