Online Prediction of Lane Change with a Hierarchical Learning-Based Approach

Liao, Xishun; Wang, Ziran; Zhao, Xuanpeng; Zhao, Zhouqiao; Han, Kyungtae; Tiwari, Prashant; Barth, Matthew; Wu, Guoyuan

doi:10.1109/icra46639.2022.9812269

Cited by 15 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Liao et al proposed an online model to predict the possibility of lane changes by surrounding vehicles. This model is a combination of two sub-modules [47]. The first sub-module uses a Long-Short Term Memory (LSTM) network, and the second sub-module uses Inverse Reinforcement Learning (IRL) to predict the trajectory of the vehicle.…”

Section: Actor-critic Networkmentioning

confidence: 99%

Multi-agent reinforcement learning for safe lane changes by connected and autonomous vehicles: A survey

Hegde

Bouroche

2024

AIC

View full text Add to dashboard Cite

Connected Autonomous vehicles (CAVs) are expected to improve the safety and efficiency of traffic by automating driving tasks. Amongst those, lane changing is particularly challenging, as it requires the vehicle to be aware of its highly-dynamic surrounding environment, make decisions, and enact them within very short time windows. As CAVs need to optimise their actions based on a large set of data collected from the environment, Reinforcement Learning (RL) has been widely used to develop CAV motion controllers. These controllers learn to make efficient and safe lane changing decisions using on-board sensors and inter-vehicle communication. This paper, first presents four overlapping fields that are key to the future of safe self-driving cars: CAVs, motion control, RL, and safe control. It then defines the requirements for a safe CAV controller. These are used firstly to compare applications of Multi-Agent Reinforcement Learning (MARL) to CAV lane change controllers. The requirements are then used to evaluate state-of-the-art safety methods used for RL-based motion controllers. The final section summarises research gaps and possible opportunities for the future development of safe MARL-based CAV motion controllers. In particular, it highlights the requirement to design MARL controllers with continuous control for lane changing. Moreover, as RL algorithms by themselves do not guarantee the level of safety required for such safety-critical applications, it offers insights and challenges to integrate safe RL methods with MARL-based CAV motion controllers.

show abstract

Section: Actor-critic Networkmentioning

confidence: 99%

Multi-agent reinforcement learning for safe lane changes by connected and autonomous vehicles: A survey

Hegde

Bouroche

2024

AIC

View full text Add to dashboard Cite

show abstract

“…By proposing a maximum causal entropy framework model, Zouzou et al [6] predicted the behavioral transformation of the vehicle in a specific situation. Meanwhile, to consider the driving behavior of the target driver, Liao et al [7] proposed a learning feedforward prediction algorithm that considers the vehicle interaction process. Wang et al [8] added semantic rewards based on adversarial inverse reinforcement learning AIRL, enabling the algorithmic framework to adapt to more challenging decision-making tasks.…”

Section: Introductionmentioning

confidence: 99%

Research on trajectory prediction of vehicle lane change for autonomous driving based on inverse reinforcement learning

Zhan,

Fan,

Jin

2024

Seventh International Conference on Traffic Engineering and Transportation System (ICTETS 2023)

View full text Add to dashboard Cite

Autonomous vehicles improve the safety and efficiency of vehicles in complex traffic scenarios through autonomous decision-making intelligence technology. To address the requirements of the self-driving vehicle lane change scenario for the accuracy of vehicle lane change trajectory prediction, in this paper, we propose a lane change trajectory prediction method for self-driving vehicles based on inverse reinforcement learning. We model the inverse reinforcement learning process through a maximum entropy mechanism to learn the optimal reward function that infers the potential end targets during the vehicle lane change. This reward model is used to construct the optimal policy that can be sampled for planning in the grid world. Conditioned on the sequence of state actions sampled by this maximum entropy policy, we generate vehicle lane change prediction trajectories. We conduct training experiments on lane change scenario data from the publicly available nuScenes dataset for autonomous driving, which shows that our method can meet the vehicle lane change requirements in real scenarios and validate the accuracy and reasonableness of the lane change trajectories.

show abstract

Social Behavior-Aware Driving Intention Detection Using Spatio-Temporal Attention Network

Shen

Pan

et al. 2023

Communications in Computer and Information Science

View full text Add to dashboard Cite

Online Prediction of Lane Change with a Hierarchical Learning-Based Approach

Cited by 15 publications

References 19 publications

Multi-agent reinforcement learning for safe lane changes by connected and autonomous vehicles: A survey

Multi-agent reinforcement learning for safe lane changes by connected and autonomous vehicles: A survey

Research on trajectory prediction of vehicle lane change for autonomous driving based on inverse reinforcement learning

Social Behavior-Aware Driving Intention Detection Using Spatio-Temporal Attention Network

Contact Info

Product

Resources

About