Code and data underlying the publication: Social-aware Planning and Control for Automated Vehicles Based on Driving Risk Field and Model Predictive Contouring Control: Driving through Roundabouts as a Case Study

This is the code and data related to the publication:L. Zhang, Y. Dong, H. Farah and B. van Arem, "Social-Aware Planning and Control for Automated Vehicles Based on Driving Risk Field and Model Predictive Contouring Control: Driving Through Roundabouts as a Case Study," 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Honolulu, Oahu, HI, USA, 2023, pp. 3297-3304, doi: 10.1109/SMC53992.2023.10394462.<br>https://doi.org/10.1109/SMC53992.2023.10394462<br>keywords: {Trajectory planning;Predictive models;Prediction algorithms;Robustness;Planning;Predictive control;Vehicles;Automated vehicles;Planning and control;Social-aware driving;Roundabouts;Driving Risk Field;Model Predictive Contouring Control},<br>The implementation is based on Python and Highway_env simulation environment https://github.com/Farama-Foundation/HighwayEnv~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br>The gradual deployment of automated vehicles (AVs) results in mixed traffic where AVs will interact with human-driven vehicles (HDVs). Thus, social-aware motion planning and control while considering interactions with HDVs on the road is critical for AVs’ deployment and safe driving under various manoeuvres. Previous research mostly focuses on the trajectory planning of AVs using Model Predictive Control or other relevant methods, while seldom considering the integrated planning and control of AVs altogether to simplify the whole pipeline architecture. Furthermore, there are very limited studies on social-aware driving that makes AVs understandable and expected by human drivers, and none when it comes to the challenging manoeuvre of driving through roundabouts. To fill these research gaps, this paper develops an integrated social-aware planning and control algorithm for AVs’ driving through roundabouts based on Driving Risk Field (DRF), Social Value Orientation (SVO), and Model Predictive Contouring Control (MPCC), i.e., DRF-SVO-MPCC. The proposed method is tested and verified with simulation on the open-sourced <em>highway-env</em> platform. Compared with the baseline method using purely Nonlinear Model Predictive Control, the DRF-SVO-MPCC can achieve better performance under various maneuvers of driving through roundabouts with and without surrounding HDVs.<br><br>

本数据集包含与下述学术论文相关的代码与数据：L. Zhang、Y. Dong、H. Farah与B. van Arem所著《基于驾驶风险场与模型预测轮廓控制的自动驾驶车辆社交感知规划与控制——以环岛通行为例》，发表于2023年IEEE系统、人与控制论国际会议（SMC 2023），会议举办地为美国夏威夷州瓦胡岛火奴鲁鲁，页码范围3297-3304，DOI：10.1109/SMC53992.2023.10394462，论文链接：https://doi.org/10.1109/SMC53992.2023.10394462。

关键词：轨迹规划；预测模型；预测算法；鲁棒性；规划；预测控制；车辆；自动驾驶车辆（Automated Vehicles, AVs）；规划与控制；社交感知驾驶；环岛；驾驶风险场（Driving Risk Field, DRF）；模型预测轮廓控制（Model Predictive Contouring Control, MPCC）。

本项目基于Python语言与Highway_env仿真平台开发，平台开源地址为：https://github.com/Farama-Foundation/HighwayEnv。

自动驾驶车辆（Automated Vehicles, AVs）的逐步普及催生了混合交通场景，即自动驾驶车辆需与人类驾驶车辆（Human-driven Vehicles, HDVs）进行交互。因此，在道路交互场景中兼顾社交感知的运动规划与控制，对于自动驾驶车辆的落地部署以及各类驾驶工况下的安全行车至关重要。既往研究多聚焦于采用模型预测控制或相关方法实现自动驾驶车辆的轨迹规划，却鲜有将规划与控制进行一体化设计以简化整体管线架构的研究。此外，现有针对可被人类驾驶员理解与预判的社交感知驾驶研究极为匮乏，针对环岛通行这类高难度驾驶工况的相关研究更是空白。为填补上述研究空白，本文提出了一种基于驾驶风险场（DRF）、社交价值取向（Social Value Orientation, SVO）与模型预测轮廓控制（MPCC）的一体化社交感知规划与控制算法（即DRF-SVO-MPCC框架），用于自动驾驶车辆的环岛通行场景。所提方法在开源的highway-env仿真平台上进行了测试与验证。相较于仅采用非线性模型预测控制的基线方法，DRF-SVO-MPCC在存在或不存在周边人类驾驶车辆的各类环岛通行工况下均可取得更优的性能。