Enhancing Mixed Traffic Flow Safety via Connected and Autonomous Vehicle Trajectory Planning with a Reinforcement Learning Approach

Abstract

The longitudinal trajectory planning of connected and autonomous vehicle (CAV) has been widely studied in the literature to reduce travel time or fuel consumptions. The safety impact of CAV trajectory planning to the mixed traffic flow with both CAV and human-driven vehicle (HDV), however, is not well understood yet. This study presents a reinforcement learning modeling approach, named Monte Carlo tree search-based autonomous vehicle safety algorithm, or MCTS-AVS, to optimize the safety of mixed traffic flow, on a one-lane roadway with signalized intersection control. Crash potential index (CPI) is defined to quantitively measure the safety performance of the mixed traffic flow. The CAV trajectory planning problem is firstly formulated as an optimization model; then, the solution procedure based on reinforcement learning is proposed. The tree-expansion determination module and rollout termination module are developed to identify and reduce the unnecessary tree expansion, so as to train the model more efficiently towards the desired direction. The case study results showed that the proposed algorithm was able to reduce the CPI by 76.56%, when compared with a benchmark model without any intelligence, and 12.08%, when compared with another benchmark model that the team developed earlier. These results demonstrated the satisfactory performance of the proposed algorithm in enhancing the safety of the mixed traffic flow.

A Novel Dynamic Lane-Changing Trajectory Planning Model for Automated Vehicles Based on Reinforcement Learning

“…In recent years, RL is increasingly used in vehicle control problems in ramp metering [26,27], intersection [28,29], and freeway work zone [30], in order to improve traffic conditions. RL-based control models for AVs have a positive impact on traffic safety and efficiency [25,31,32]. For lanechanging problems of AVs, most RL-based studies focus on the lane-changing decision process [33][34][35][36][37][38], and the application of RL to lane-changing trajectory planning still needs to be supplemented.…”

Research on Influencing Factors of Urban Road Traffic Casualties through Support Vector Machine

“…Some scholars focused on the interaction between autonomous driving trajectory planning and mixed traffic flow. It presented a reinforcement-learning modeling approach to optimize the safety of mixed traffic flow on a single-vehicle road controlled by a signalized intersection [6]. Government managers can develop a coordination policy based on "catch-up" to gradually increase the penetration rate of the CAV market to reduce traffic oscillation and improve road capacity [35].…”

“…Related research showed that a greater penetration rate of autonomous vehicles in traffic flow would ensure a more stable traffic flow and a lower risk [4]. Moreover, the penetration rate of autonomous driving can also optimize the average speed, delay of road traffic flow, and the safety performance of mixed traffic flow [5,6].…”

Koordine Kavşaklarda Sinyalizasyon Sistemine Etki Eden Parametrelerin İncelenmesi Vaka Durum Çalışması

“…Koridordaki ortalama hız ve duruş sayısı ölçütlerinin beraber değerlendirilmesiyle, geleneksel yaklaşımlara göre daha hassas değerlendirme yapabilen, karma yaklaşım modelleri ortaya çıkmıştır [30]. Otonom ve bağlı araç teknolojilerinin yakın gelecekte kullanılabilme ve yaygınlaşma potansiyeli nedeniyle, son zamanlarda gerçekleştirilen çalışmalarda, bu teknolojiler dikkate alınmış ve eşgüdümlü kavşak tasarımında, otonom araçların da bulunduğu karma trafik koşullarını içeren bazı modeller geliştirilmiştir [31,32,33 Trafik mühendisliğinde, birbirleriyle ilişkili olarak yönetilen kavşaklara, eşgüdümlü kavşak; ayrı olarak yönetilen (ilişkisi bulunmayan) kavşaklara ise, ayrık (izole) kavşak adı verilmektedir. Kavşakların, eşgüdümlü şekilde çalışabilmeleri için, birbirlerine yakın mesafede (en fazla 600-800 m) ve komşu olmaları gerekmektedir.…”