Post-impact motion control of road vehicles

Post-Impact Control (PIC) deals with the motion control of a motor vehicle after a ‘light impact’, one that may significantly disturb the pre-impact motion, but largely leaves the vehicle intact and in a controllable state. Typically, the initial impact is severe enough for the driver to lose control, leading to an increased risk of a highly damaging secondary event, such as road departure, rollover or collision with another obstacle. Control of vehicle post-impact motion is expected to be fully or partially automated, so that driver skill and disorientation become less of a risk factor. Also referred to as post-impact stability control (PISC), this topic has been studied for more than 20 years. However, there has been a recent impetus to revisit the topic based on a number of developments: (i) increasing numbers of motion control actuators and sensors, such as torque vectoring, semi-active suspension, steer-by-wire and brake-by-wire systems; (ii) supporting research with algorithms, such as Model Predictive Control (MPC) to improve performance, including in autonomous driving (AD); (iii) safety concerns relating to new AD systems and their ability to respond to light impacts; (iv) new capabilities in terms of high precision digital maps and sensing systems that evaluate the dynamic driving environment in real-time. Notably, new research has led to a number of recent publications in PIC, with a particular focus on the use of MPC. Vehicle system dynamics in a post-impact state may involve large initial yaw rate and body sideslip angle, much greater than encountered in normal driving, and hence unique features are affecting the need for collision sensing, motion estimation and control methods. PIC may include several levels of sophistication, such as post-impact braking (PIB), which extends the range of autonomous braking, post-impact stability control (PISC) which extends both the dynamic range and functionality of ESC, up to more advanced levels where full automation brings the impacted vehicle to a safe condition via a so-called minimum risk manoeuvre. Future research in PIC is required to fully map such risk factors in real-time, and deliver autonomous post-crash responses where all motion variables are controlled appropriately in very short timescales and with tyre adhesion and stability limits as critical factors. By reviewing existing research on post-impact control, this paper proposes metrics and algorithms appropriate to real-world PIC implementation, including a framework for system design, testing and validation.