Session: Poster Session

Paper Number: 148264

148264 - Optimization-Based Predictive Controller to Follow a Trajectory With Multiple Turns 

Autonomous vehicles have been developed for human comfort and convenience. These vehicles need precisely designed control systems to increase their reliability and accuracy. Conventional feedback controllers utilize precomputed controls law and lack the capability to predict and optimize the trajectory following. This makes them less efficient in the case where vehicle and environment constraint necessitates higher forms of control. To address this, model predictive control, suggests use of a mathematical model of a system to make predictions about its future behavior. In this study, we implement the model predictive control as a control strategy for our vehicle which follows a trajectory with two turns. We use the kinematic bicycle model to simplify the motion of the vehicle. The governing equations are discretized using the forward Euler method. This process involves approximating the continuous-time dynamics of the model with an Euler discrete-time representation that is used by the MPC controller. The discretized form of the kinematic bicycle model then is used to predict the future behavior of the system and to compute optimal control inputs for the vehicle. To model the appropriate trajectory, a third-order polynomial regression between each of the target points is used. Our results reveal that the model is capable of accurately following the trajectory and the corresponding optimized state results are reported. 

Presenting Author: Shila Alizadehghobadi University of California Merced

Presenting Author Biography: Shila Alizadehghobadi is currently a PhD student in control and dynamical systems at the University of California, Merced. In 2022, she was awarded a Eugene Cota Robles Fellowship to do her PhD studies. In 2023, she was a visiting student at University of California, Berkeley, and the same year she presented multiple talks on optimization based control systems. She was also selected as finalist in Falling Walls Lab competition in 2023. Throughout the years, she has contributed to diversity, equity and inclusion (DEI), and she is currently serving as co-president of Graduate Women of Engineering (GWE) at UC Berkeley.

Authors:

Shila Alizadehghobadi University of California Merced