A nonlinear and optimal control simplification technique for real-time lane-changing maneuvers

The automotive industry is actively seeking novel ideas to improve vehicles’ safety and efficiency through real-time solutions for relating complex optimization problems. The present study aims to diminish the computing burden of a complicated optimal lane-changing procedure as well as a nonlinear controller that is responsible for lateral vehicle dynamics. As such, some mathematical functions for trajectories are examined to find the most appropriate one for lateral movement while considering safety and efficiency through designing a constrained optimization problem that determines the endpoint of the lane-changing maneuver. Next, the aforementioned problem is explored to discover a simple correlation instead of the huge-scale optimum-finding problem. As a result, the endpoint of the procedure is indicated by a linear equation. Consequently, a Non-linear Model Predictive Controller (NMPC) is designed to control vehicle dynamics in two directions. Then, a simple and novel sinusoidal function with varying amplitude and period is proposed. This function appears to be a steering angle profile which results in an optimal and safe lane change process without going through cumbersome procedure. In other words, the constrained optimization problem and nonlinear MPC will be replaced by a simple sinusoidal function. Also, the effectiveness of the scheme is validated in Siemens PreScan ® and MATLAB/SIMULINK.