KEYWORDS – haptic shared control (HSC), model predictive control (MPC), human-machine interaction, handling limits, safety envelope. ABSTRACT OBJECTIVE: Leveraging recent developments in sensing and actuation technologies, the authors propose a novel predictive control framework for understeer prevention using haptic feedback provided to the driver directly on the steering wheel. The objective of the proposed system is to intuitively alert the driver about incoming front tire saturation limits before reaching them as well as to provide support in handling vehicle understeer correctly. Such scenarios are relevant due to the high likelihood to cause accidents, as regular drivers are not prepared to handle these conditions safely. METHODOLOGY: A model predictive control (MPC) method is used to predict the future vehicle lateral velocity, yaw rate and steering angle over an imminent time horizon using a bicycle vehicle model and a nonlinear brush tire model. A safe steering envelope is defined for every timestep of the prediction horizon. This envelope represents the saturation limits of the vehicle’s front tires; operation within this envelope ensures that the vehicle remains within its handling limits. In case the predicted steering angle violates the safe envelope, haptic feedback is provided on the steering wheel in the form of an increasingly opposing torque with vibrations. The haptic driver support system is compared with manual steering in a path-tracking driving study involving an obstacle avoidance maneuver in the middle of a turn. The study was realized on a high-fidelity driving simulator with 32 participants. RESULTS: The participants were divided into three groups based on their driving experience (a. expert, b. regular, c. novice). During the obstacle avoidance maneuver, the steering behavior of all three participant groups is positively influenced by the haptic driver support, which is objectively represented by a smaller peak steering angle with respect to manual steering (Δ?_???,? = −8.3∘, Δ?_???,? = −63.9∘, Δ?_???,? = −52.6∘). This results in a lower peak slip angle at the front tires (Δ?_???,? = −0.0089 ???, Δ?_???,? = −0.0553 ???, Δ?_???,? = −0.0475 ???) and a smaller peak lateral deviation in case of the regular and the novice drivers (Δ?_? = +0.038 ?, Δ?_? = −1.070 ?, Δ?_? = −0.354 ?). Furthermore, subjective evaluation indicates strong acceptance by the participants, with 24 out of 32 participants that preferred to drive with the haptic support during the experiment and 26 out of 32 participants that expressed their interest in having such a system installed in their own car. LIMITATIONS: The haptic driver support system assumes constant longitudinal velocity of the vehicle. Furthermore, in order to compute the safe steering envelope, knowledge about the tire-road friction coefficient is required. NOVELTY: To the best of the author’s knowledge, the proposed system is a first haptic driver support system with predictive capabilities for the vehicle handling limits use case. CONCLUSION: The results obtained during the human-in-the-loop experiments using the driving simulator show the potential to increase safety margins and avoid (potentially) fatal accidents. We demonstrate the positive impact of the haptic driver support system on the driver’s behavior which results in improved vehicle handling near the saturation limit of the front tires, both objectively and subjectively. Further research is needed to adapt the driver support system to varying speeds. Additionally, an on-road study with a real vehicle should be considered to validate the proposed system.

Mr. Kazimierz Dokurno, Student, Delft University of Technology