Robust Lane Change Decision Making for Autonomous Vehicles: An Observation Adversarial Reinforcement Learning Approach

He, Xiangkun; Yang, Haohan; Hu, Zhongxu; Lv, Chen

doi:10.1109/tiv.2022.3165178

Cited by 96 publications

(18 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, it can be clearly seen from Figure 13 that the performance of the proposed algorithm is better than the DQN algorithm without multi-layer safety lane changing strategy in the training process. This means that FIGURE 14 The safety reward in lane changing moment of three traffic scenarios. the proposed ML-SLCS method has higher efficiency in the training process.…”

Section: Figure 13mentioning

confidence: 99%

“…They adapt the hierarchical double deep Q‐learning (h‐DDQN) algorithm to make the framework robust and use the hybrid reward mechanism and reward‐driven exploration to improve the sample efficiency. To ensure safety under perception uncertainty, a novel observation adversarial reinforcement learning approach for robust lane changing decision making in [14]. It can not only enhance the platoon performance but also improve the robustness of lane changing policies against adversarial observation perturbations.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep Q‐network based multi‐layer safety lane changing strategy for vehicle platoon

Zhang,

Yan,

Zuo

2023

IET Intelligent Trans Sys

View full text Add to dashboard Cite

The vehicle platoon lane changing is significant for alleviating road congestion and diminishing transportation energy consumption. However, the lane changing strategy for a group of vehicles is still a great challenge in this field. This paper investigates the vehicle platoon lane changing problems, in which the safety and efficiency in the lane changing procedure are both taken into consideration. Since the safety of the platoon lane changing would be affected by the lane changing gap and the length of the platoon, a novel platoon lane changing strategy is proposed by using the deep Q‐network. In detail, the proposed platoon lane changing strategy contains two layers, where the first one is a decision layer and the other one is the verification layer. In the decision layer, the deep Q‐network is employed to improve the lane changing efficiency. Then, the verification layer is presented to enhance the platoon lane changing safety. In final, some typical platoon lane changing scenarios are provided in an existing ramp containing a vehicle platoon and some random vehicles. The related numerical simulations are conducted to validate the feasibility and effectiveness of the proposed approaches.

show abstract

Section: Figure 13mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Deep Q‐network based multi‐layer safety lane changing strategy for vehicle platoon

Zhang,

Yan,

Zuo

2023

IET Intelligent Trans Sys

View full text Add to dashboard Cite

show abstract

“…When predicting driver intention, lane change is the most commonly encountered intention. Here, the target includes the ego-vehicle [136,143] and surrounding vehicles [144]. The prediction of surrounding vehicles utilizes trajectory information, obtained based on the GPS and Internet of Vehicles(IoV) technology, as the input to infer the intention.…”

Section: Driver Intention Inference and Predictionmentioning

confidence: 99%

Review and Perspectives on Driver Digital Twin and Its Enabling Technologies for Intelligent Vehicles

Lou

Xing

et al. 2022

IEEE Trans. Intell. Veh.

Self Cite

View full text Add to dashboard Cite

Digital Twin (DT) is an emerging technology and has been introduced into intelligent driving and transportation systems to digitize and synergize connected automated vehicles. However, existing studies focus on the design of the automated vehicle, whereas the digitization of the human driver, who plays an important role in driving, is largely ignored. Furthermore, previous driver-related tasks are limited to specific scenarios and have limited applicability. Thus, a novel concept of a driver digital twin (DDT) is proposed in this study to bridge the gap between existing automated driving systems and fully digitized ones and aid in the development of a complete driving human cyber-physical system (H-CPS). This concept is essential for constructing a harmonious human-centric intelligent driving system that considers the proactivity and sensitivity of the human driver. The primary characteristics of the DDT include multimodal state fusion, personalized modeling, and time variance. Compared with the original DT, the proposed DDT emphasizes on internal personality and capability with respect to the external physiological-level state. This study systematically illustrates the DDT and outlines its key enabling aspects. The related technologies are comprehensively reviewed and discussed with a view to improving them by leveraging the DDT. In addition, the potential applications and unsettled challenges are considered. This study aims to provide fundamental theoretical support to researchers in determining the future scope of the DDT system.

show abstract

“…With its rapid development, driverless technology is now widely employed in various mine transportation scenarios [1,2]. One area in which this technology can be applied is in monorail cranes, which, when used in mine auxiliary transport systems, play an invaluable role in complicated mining settings because of their excellent transport and * Author to whom any correspondence should be addressed.…”

Section: Introductionmentioning

confidence: 99%

DFFRLS-FAUKF: accurate and reliable monorail longitudinal slope identification method

Liu,

Li,

Zheng

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Monorail cranes have always played an important role in mine auxiliary transportation systems owing to their excellent transportation performance and are therefore a desirable area in which to apply driverless technologies. However, the low-accuracy recognition of monorail track slopes and the poor reliability of recognition results make it difficult and dangerous to implement fully driverless monorail cranes. Aiming to solve these problems, a method for the accurate identification of longitudinal monorail slopes based on the use of a dynamic forgetting factor for recursive least squares (DFFRLS) and a fuzzy adaptive unscented Kalman filter (FAUKF) is proposed. First, acquired acceleration and velocity data are pre-processed using a rolling window. Second, the real-time longitudinal track-curvature value is calculated using the DFFRLS algorithm with the processed data and an established track-curvature model. Finally, based on existing track-curvature values, dynamic recognition of the monorail track slope is realised using the FAUKF algorithm with a fuzzy control factor, improving the accuracy of track gradient recognition. Experiments show that the DFFRLS-FAUKF algorithm improves the accuracy of track-slope recognition by up to 21.26% and 33.93% on average compared with that of DFFRLS with an adaptive extended Kalman filter (DFFRLS-AEKF) or an adaptive unscented Kalman filter (DFFRLS-AUKF).

show abstract

Robust Lane Change Decision Making for Autonomous Vehicles: An Observation Adversarial Reinforcement Learning Approach

Cited by 96 publications

References 35 publications

Deep Q‐network based multi‐layer safety lane changing strategy for vehicle platoon

Deep Q‐network based multi‐layer safety lane changing strategy for vehicle platoon

Review and Perspectives on Driver Digital Twin and Its Enabling Technologies for Intelligent Vehicles

DFFRLS-FAUKF: accurate and reliable monorail longitudinal slope identification method

Contact Info

Product

Resources

About