Eco-driving Strategy for Connected Automated Vehicles in Mixed Traffic Flow
Hongjie Liu,Tengfei Yuan,Xiaoqing Zeng,Kaiyi Guo,Yizeng Wang,Yanghui Mo,Hongzhe Xu
DOI: https://doi.org/10.1016/j.physa.2023.129388
IF: 3.778
2024-01-01
Physica A Statistical Mechanics and its Applications
Abstract:Mixed traffic flow is a prevalent phenomenon in the trend of connected automated vehicles (CAVs), where a diverse set of road users, including cars, motorcycles, bicycles, pedestrians, and even animals, share the road infrastructure. This coexistence poses a range of challenges, not limited to traffic safety, efficiency, and environmental sustainability. Compared with the traditional traffic streams, the controllability of connected and automated vehicles within mixed traffic offers new possibilities for eco-driving. As CAV technologies continue to flourish, this study explores the imperative of constructing eco-roads within a mixed traffic framework and optimizing eco- driving strategies to enhance vehicle energy efficiency and reduce emissions. We extended the concept of mixed traffic flow to incorporate scenarios involving animal crossings, introducing an eco-road-based green mixed traffic model. Analyzing the driving behaviors of both autonomous and manually-driven vehicles within a vehicular network ecosystem, we proposed an eco- road driving model that includes vehicle-following and lane-changing behaviors. From the perspective of dynamic programming, we conducted a discrete analysis to create an energy- saving driving model apt for mixed traffic conditions, with Q-Learning serving as the optimal solver. We further validated our theoretical framework through simulations conducted on eco- roads in Shanghai, taking into account the inherent risks brought about by the crossing of wildlife. Our empirical results indicated that the recommended energy-saving strategies could potentially reduce fuel consumption by 6–11%. Interestingly, the energy-saving effects are amplified with an increasing density of networked autonomous vehicles (CAVs) within the mixed traffic environment. Our findings strengthen the feasibility of our model and the efficacy of our algorithmic approach, confirming that the described driving strategies hold great promise for significantly improving energy efficiency in the domain of connected vehicles under the premise of ensuring the safety of driving and wildlife.