A model predictive control-based lane merging strategy for autonomous vehicles
A model predictive control-based lane merging strategy for autonomous vehicles
复制标题
基于模型预测控制的自动驾驶车辆车道合并策略
DOI：
10.1109/ivs.2019.8814171
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Suzuki Tatsuya
中科院分区：
文献类型：
--
作者：
Tran Anh Tuan;Kawaguchi Masato;Okuda Hiroyuki;Suzuki Tatsuya
本文提出了一种模型预测控制器来执行车道合并任务。我们考虑一个分层的控制结构，其中包括一个内部的控制回路和一个外部的。内环是一个自适应巡航控制器，该控制器在满足约束条件的情况下，向自主车辆发出加速命令，以跟随指定的速度并与前车保持相对距离。外环是确定自主车辆应该跟随的主车道上的车辆，以最小化主车道上人类驾驶员决策的熵。仿真结果验证了所提出的控制器在不确定性和建模误差的情况下能够完成车道合并任务。
This paper proposes a model predictive controller to perform the lane merging task. We consider a hierarchical control structure which consists of an inner control loop and an outer one. The inner loop is an adaptive cruise controller which gives the acceleration command to the autonomous vehicle to follow the designated speed and keep a relative distance with the preceding vehicle while satisfying constraints. The outer loop is to determine the vehicle in the main lane that the autonomous one should follow to minimize the entropy in the decision making of the human drivers in the main lane. It is verified in the simulation that the proposed controller can complete the lane merging task under uncertainties and modeling errors.