A Model Predictive Controller for Automated Lane Change Maneuver with Multi-Constraints

Mingliang Yang,Kun Jiang,Weiguang Yu,Shengjie Kou,Diange Yang
DOI: https://doi.org/10.1061/9780784483565.245
2021-01-01
Abstract:Automated lane change is a challenging task due to simultaneous longitudinal and lateral vehicle control in high-speed dynamic environments. To guarantee safe and efficient lane change, a model predictive controller considering multi-constraints is presented. Safe driving space for the predictive driving environment and physical vehicle limitations is considered as a hard constraint to ensure safety. Inter-vehicle car-following distance is considered as a soft constraint to improve traffic efficiency. The algorithm solves a convex quadratic program to obtain the optimal control input. Simulation with CARLA simulator proves that the lane changing and car-following model with the constraints improves driving comfort and lane change efficiency. Combined with lane changing safety distance constraint and car following safety distance constraint, this algorithm considers dynamic characteristics of vehicles, solves the problem of no solution of trajectory planning of other models, and realizes the adaptive adjustment of lane changing safety and efficiency in different scenarios.
What problem does this paper attempt to address?