Vehicle optimal road departure prevention via model predictive control

Abstract

This article addresses the problem of road departure prevention using integrated brake control. The scenario\ud considered is when a high speed vehicle leaves the highway on a curve and enters the shoulder or another lane,\ud due to excessive speed, or where the friction of the road drops due to adverse weather conditions. In such a scenario,\ud the vehicle speed is too high for the available tyre-road friction and road departure is inevitable; however, its effect can\ud be minimized with an optimal braking strategy. To achieve online implementation, the task is formulated as a receding\ud horizon optimization problem and solved in a linear model predictive control (MPC) framework. In this formulation, a\ud nonlinear tire model is adopted in order to work properly at the friction limits. The optimization results are close to\ud those obtained previously using a particle model optimization, PPR, coupled to a control algorithm, MHA, specifically\ud designed to operate at the vehicle friction limits. This shows the MPC formulation may equally be effective for vehicle\ud control at the friction limits. The major difference here, compared to the earlier PPR/MHA control formulation, is that\ud the proposed MPC strategy directly generates an optimal brake sequence, while PPR provides an optimal reference\ud first, then MHA responds to the reference to give closed-loop actuator control. The presented MPC approach has the\ud potential to be used in futur

Unified decision-making and control for highway collision avoidance using active front steer and individual wheel torque control

“…The collision avoidance controller is to make use of the 3-degree-of-freedom (DOF) planar two-track vehicle model shown in Figure 2 [24], where v x , v y , r are the longitudinal velocity, lateral velocity and yaw rate in the vehicle body axis system.…”

“…To give an explicit formula for the lateral tire force, a friction ellipse relationship between peak longitudinal and lateral tire forces is assumed, as illustrated in Figure 3 [24]. For each tire, the equation of the friction ellipse is given by: where the subscript * * denotes f l, f r, rl, rr.…”

“…In this study, the magic formula parameters B, C, D, E for the tire model (4) are fitted with the least-squares estimation (LSE) method, using the CarSim tire data from the tire '215/55 R17' [26]. The obtained parameters are shown in Table 1 [24]. According to Table 1, the parameter E is quite small for all vertical loads, so for simplicity we use the approximation E = 0.…”

“…The elliptical relationship between longitudinal and lateral tire forces at different slip angles[24].…”

Compensatory model predictive control for post-impact trajectory tracking via active front steering and differential torque vectoring

“…It is essential that smooth actuation entails a quicker control response, because the earlier the controller intervenes, the quicker tracking errors of the post-impact vehicle can be converged to the desired zero. Quite different from the conventional MPC controllers, 28,29 a major aspect of the proposed control design in this paper lies in the specific analysis from a physical perspective. To the best of our knowledge, few research papers show the significance of the compensatory MPC on front steering and differential driving torques when post-impact vehicle states need to be manipulated.…”