A Model-Based LQR Control of an Obstacle Avoidance Maneuver of a Self-Driving Car

The safe and effective computer control of different types of autonomous vehicles is becoming an increasingly important task in modern Cyber-Physical Systems (CPS) and in future transportation solutions of the Smart Cities. Paper discusses the linearized mechanical model and Linear Quadratic Regulator (LQR) control mathematics of a Self-Driving Car (SDC) of 1200 kg weight class. We demonstrate in the Simulink/MATLAB environment the applicability of this cost functions based advanced optimized control approach for reliable performing of a two-step obstacle avoidance maneuver. Study is related to improvement of control algorithms of a real autonomous shuttle ISEAUTO operating in the campus of Tallinn University of Technology.