Adaptive optimal control allocation using Lagrangian neural networks for stability control of a 4WS–4WD electric vehicle

Abstract

This study involves a layered vehicle dynamics control system, which is composed of an adaptive optimal control allocation method using Lagrangian neural networks for optimal distribution of tyre forces and the sliding mode yaw moment observer for robust control of yaw dynamics. The proposed optimal control allocation method eliminates the requirement of solving optimization problem in every time step and it is a convergent and stability guaranteed solution for the optimal tyre force distribution problem. The aim in the sliding mode yaw moment observer is to force the vehicle to track a reference vehicle dynamic behaviour by estimating the equivalent input extended disturbance, which is the required stabilizing virtual yaw moment. The proposed layered stability control scheme has been tested on a four-wheel drive-four-wheel steer electric Fiat Doblo Van, which is modelled in CarSim. Both the sliding mode disturbance observer and the optimal control allocation methods are the first known applications to the stability control problem of road vehicles.

Optimal Distribution Control Of Non‐Linear Tire Force Of Electric Vehicles With In‐Wheel Motors

“…7(a) and 7(b), respectively. For the purpose of comparison, the simulation result of the linear optimal control allocation method using equations (12) and (13) is also presented. The simulation performance is also presented when no optimal tire force distribution method is applied.…”

“…Suzuki et al used the simple linear relations between the steering angle, driving torque, and side force F si or traction or brake force F ti as shown in equations (12) and (13) [14]…”

“…where Δδ i is the adjusted steering angle of each wheel, which is added into the distributed steering angle δ i calculated from the linear model (12). a i represents the actual slip angle of the individual wheel obtained from the vehicle dynamics model and a di represents the desired slip angle of the individual wheel when considering the non-linear tire characteristics.…”

“…used a virtual control law determined offline to implement the CA method . The adaptive CA method was also proposed to save computational time, since the CA problem can be solved dynamically without needing to find the optimal solution at each sampling instant . The main problem with the above‐mentioned methods is that they need to solve a very complex optimization problem.…”

A bilateral control scheme for vehicle steer-by-wire system with road feel and steering controller design

“…The X-by-wire (XBW) technology has received much attention since it provides the potential benefit for enhancing vehicle performance (Liu et al, 2017a; Sun et al, 2015), improving comfort (Do et al, 2014), stability (Yao et al, 2014a), maneuverability, and fully integrated vehicle dynamic control (Demirci, 2013). A steer-by-wire (SBW) system is such a system where there is no mechanical connection between the steering wheel and the front road wheels (Wei et al, 2016; Yamaguchi and Murakami, 2009), whilst the electric actuators are deployed to control the steering wheel and the front wheel independently (Wu et al, 2016).…”