Lane changing (LC) is a typical driving task of connected and autonomous vehicles (CAV). In a typical LC event, CAV need to follow certain trajectory and enters the gap at target lane. While there exists many LC planning models, gap feasibility, which should be a preposition stage of LC planning, is rarely investigated. Furthermore, if the neighboring gap is not feasible, CAV needs to accelerate/decelerate to accommodate downstream/upstream gaps. This LC process with adaption then include an accelerate/decelerate part and a LC part. The terminal state of the first part is called the lane changing initiation state (LCIS), i.e. the initiation state of the second part. The downstream/upstream gap feasibility analysis needs to determine the LCIS, a state set rather than a single state. The above gap feasibility problem is more complex and current methods cannot solve this problem. This research proposes a forward-backward reachability model to address this issue. A forward-backward vehicle dynamics model is constructed for this purpose. The Pontryagin Maximal Principle (PMP) is employed to determine the speed curve of the acceleration/deceleration part. The proposed method is tested using a numerical example, which proves the usefulness of the model.
Gap feasibility with and without adaption for autonomous vehicle's lane changing: a forward-backward reachability approach
08.10.2022
519691 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
The Vehicle’s Lane-Changing Rules of On-Ramp System
Trans Tech Publications | 2013
|Method and Apparatus for Vehicle's Exiting from Reversible Lane
Europäisches Patentamt | 2023
|