Fundamental approaches in modeling the control of a car by a driver are to be reviewed with careful consideration. The context of the work is to analyze how the car needs advanced development while driving. The previous research on the appliance of optimal linear preview control to aspects of driving road vehicles is extended. This prior research treated the tracking of a roadway by a vehicle and driver at constant forward speed and therefore the tracking of a speed demand while running straight. The 2 previously separate problems are now combined, so that longitudinal and lateral path demands are considered in parallel.

A new feature is that low-pass filters are included within the driver modeling, to represent driver bandwidth limitations. This feature enables the finding of the influence of the driver’s control bandwidth on the optimal strategies and the closed-loop performance, by way of frequency-response calculations. A replacement optimal preview control toolbox is used. Simulations of the virtual driver-controlled car are shown to demonstrate the closed-loop following longitudinal and lateral position demands.

The dynamics of a racer are often described as follows:

Primary interest from an impact viewpoint is within the longitudinal and lateral dynamics. Longitudinal control is especially exercised by the throttle and brake controls, while lateral control comes from the steering mechanism. It’s conventional to consider the throttle displacement, the pedal pressure, and therefore the wheel displacement as being the many control inputs. The longitudinal dynamics are simpler than the lateral dynamics, but both must be considered together in treating the overall maneuvering problem.

The dynamic characteristics of the race-car vary strongly with speed. Within an operating range around the straight-running state, a linear representation of a (good) car is predicted to be accurate. As for maneuvering severity increases, tire shear forces saturate during a smooth and progressive manner. Almost saturation, shear forces depend upon the frictional coupling between the tires and therefore the ground and longitudinal and lateral forces compete for the available friction. Controls derived assuming linearity are expected to figure well for gentle maneuvering and not so well for the limit operation.

A racing driver is often expected to understand the dynamics of his/her car perfectly. The optimally of controls is significant. Robustness isn’t so important. Neglecting practical issues concerning computing time, standard Non-Linear Model Predictive Control (NMPC) theory would offer virtual racing driver designs straightforwardly.

Such theory involves the parameterization of the trail before the vehicle and of the control history within the preview/control horizon. Then, at each computational step, a high-dimension nonlinear optimization problem has got to be solved online. Convergence to the worldwide optimum isn’t guaranteed and results are largely hidden from view. From any solution obtained, only the primary step is employed for control. The method is slow and lavish. Compromise between the accuracy of solution and speed of computation is important.