Published on Fri Jul 14 2017

Freeway Merging in Congested Traffic based on Multipolicy Decision Making with Passive Actor Critic

Tomoki Nishi, Prashant Doshi, Danil Prokhorov

Freeway merging in congested traffic is a significant challenge toward fully automated driving. Merging vehicles need to decide not only how to merge into a spot, but also where to merge. We present a method for the freeway merging based on multi-policy decision making.

0
0
0
Abstract

Freeway merging in congested traffic is a significant challenge toward fully automated driving. Merging vehicles need to decide not only how to merge into a spot, but also where to merge. We present a method for the freeway merging based on multi-policy decision making with a reinforcement learning method called {\em passive actor-critic} (pAC), which learns with less knowledge of the system and without active exploration. The method selects a merging spot candidate by using the state value learned with pAC. We evaluate our method using real traffic data. Our experiments show that pAC achieves 92\% success rate to merge into a freeway, which is comparable to human decision making.

Tue Jan 21 2020
Machine Learning
Cooperative Highway Work Zone Merge Control based on Reinforcement Learning in A Connected and Automated Environment
Work zone merge control is critical for improving work zone safety and capacity. The proposed method assumes that all vehicles are fully automated and connected. It inserts two metering zones in the open lane to make space for merging vehicles in the closed lane.
0
0
0
Mon Mar 22 2021
Artificial Intelligence
Learning to Robustly Negotiate Bi-Directional Lane Usage in High-Conflict Driving Scenarios
0
0
0
Mon Mar 08 2021
Machine Learning
Decision-Making under On-Ramp merge Scenarios by Distributional Soft Actor-Critic Algorithm
0
0
0
Mon Dec 09 2019
Artificial Intelligence
Intelligent Coordination among Multiple Traffic Intersections Using Multi-Agent Reinforcement Learning
We use Asynchronous Advantage Actor Critic (A3C) for implementing an AI agent that optimize flow of traffic across a single intersection. We then extend it to multiple intersections by considering a multi-agent setting. We observe that (1) & (2) leads to a reduction in
0
0
0
Mon Mar 02 2020
Artificial Intelligence
Dynamic Queue-Jump Lane for Emergency Vehicles under Partially Connected Settings: A Multi-Agent Deep Reinforcement Learning Approach
Emergency vehicle (EMV) service is a key function of cities and is challenging due to urban traffic congestion. The main reason behind delay is the lack of communication and cooperation between vehicles. In this paper, we study the improvement of EMV service under V2Xconnectivity.
0
0
0
Wed Jun 26 2019
Artificial Intelligence
Cooperation-Aware Reinforcement Learning for Merging in Dense Traffic
Decision making in dense traffic can be challenging for autonomous vehicles. Human drivers leverage the cooperation of other drivers to avoid deadlock situations. Decision algorithms must reason about the interaction with other drivers.
0
0
0