Automated Lane Change Decision Making using Deep Reinforcement Learning in Dynamic and Uncertain Highway Environment

Abstract

Autonomous lane changing is a critical feature for advanced autonomous driving systems, that involves several challenges such as uncertainty in other driver's behaviors and the trade-off between safety and agility. In this work, we develop a novel simulation environment that emulates these challenges and train a deep reinforcement learning agent that yields consistent performance in a variety of dynamic and uncertain traffic scenarios. Results show that the proposed data-driven approach performs significantly better in noisy environments compared to methods that rely solely on heuristics. * These authors contributed equally to this work 1 A. Alizadeh is with

Survey of Deep Reinforcement Learning for Motion Planning of Autonomous Vehicles

“…9). The same approach can be seen in [55], though extending the number of traced objects to nine. These researches lack lateral information, though, in [54], the lateral positions and speeds are also involved in the input vector resulting in a 6x(dx, dy, dvx, dvy) structure, logically representing longitudinal and lateral distance, and speed differences to the ego, respectively.…”

“…There are lighter approaches, where the episode terminates with failure before the accident occurred, with examples of having a too high tangent angle to the track or reaching too close to other participants. These "before accident" terminations speed up the training by bringing the information of failure forward in time, though their design needs caution [55].…”

“…Making decisions on the behavioral layer consists of at least three discrete actions: Keeping current lane, Change to the left, and Change to the right, as can be seen in [55]. In this paper, the authors used the ground truth information about the ego vehicle's speed and lane position, and the relative position and speed of the eight surrounding vehicles as the observation space.…”

“…Some papers use "ground truth" environment representations or "ideal" sensor models, and only a few articles utilize sensor noise. On the one hand, transferring the knowledge acquired from ideal observations to real-world application poses several feasibility questions [122], on the other hand, using noisy or erroneous models could lead to actually more robust agents, as stated in [55].…”

Self-Adaptive Motion Prediction-Based Proactive Motion Planning for Autonomous Driving in Urban Environments

“…Bouton et al used Reinforcement Learning (RL) with linear temporal logic to determine the driving policy at unsignalized intersections [46]. In addition, RL was proposed to construct a lane change decision algorithm [47,48].…”