Model Predictive Controller for Vehicle lane change and merge

Association - Carnegie Mellon University

Motivation:

The motivation for this project stems from the highways in the USA. When merging onto a highway from an exit, vehicles encounter heavy traffic, making it challeging for the ego vehicle to merge optimally.

Application:

This work aims to develop a model and control structure for autonomously merging a vehicle into traffic lanes where certain speed and spacing constraints are assumed. The approach involves generating a simple model of the system dynamics, and utilizing model predictive control (MPC) techniques to provide control effort that allows for lane changing under quantifiable constraints

overview figure
Figure 1:System Architecture.