A Long-Short-Term Mixed-Integer Formulation for Highway Motion Planning
Welcome to the project page of “A Long-Short-Term Mixed-Integer Formulation for Highway Motion Planning”. This work is related to an IEEE journal submission by Rudolf Reiter, Armin Nurkanovic, Daniele Bernardini, Moritz Diehl and Alberto Bemporad.
Deterministic Planning - Comparison of Planners
The following simulations show, how the LSTMP, the A-star planner and the MIPDM plan and execute lane changes in deterministic closed-loop simulations. Each simulation consists of randomized surrounding vehicles with different speeds. Surrounding vehilces break, if they are close to a leading vehicle. The the LSTMP uses a safe short-term trajectory (black) to plan the first lane change and uses transition points in the spatio-temporal space for long-term planning (green). The full planning problem is formualated as MIQP and solved in approximately 50ms.
Scenario 1:
LSTMP:
A-star:
MIPDM (N=15):
MIPDM (N=20):
Scenario 2:
LSTMP:
A-star:
MIPDM (N=15):
MIPDM (N=20):
Interactive SUMO Simulations
In the following, interactive sumo simulations based on CommonRoad scenarios are shown. The planner aims to safely reach the left-most lane.
Scenario US101-22_1_I-1-1: