A Convex Optimization Based Autonomous Intersection Control Strategy in Vehicular Cyber-Physical Systems Penglin Dai ∗ , Kai Liu ∗ , Qingfeng Zhuge ∗ , Edwin H.-M. Sha ∗ , Victor Chung Sing Lee, † and Sang Hyuk Son ‡ ∗ College of Computer Science, Chongqing University, Chongqing, 400044, China Email: {penglindai, liukai0807, qfzhuge, edwinsha}@cqu.edu.cn † Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong Email: csvlee@cityu.edu.hk ‡ Information and Communication Engineering, Daegu Gyeongbuk Institute of Science and Technology, Daegu 711-873, Korea Email: son@dgist.ac.kr Abstract—Vehicular Cyber Physical Systems (VCPS) are envi- sioned to enable novel approaches to managing and controlling future road traffic intersections. Previous efforts on autonomous intersection control (AIC) mainly focused on collisions avoid- ance and traffic efficiency, without considering travel experience from passengers’ perspective. In this work, we design a convex optimization based AIC mechanism. In particular, we design the metric of smoothness with the objective of quantitatively capturing the quality of travel experience and transform the objective into a convex function. In addition, we linearize colli- sion avoidance constraints by designing a schedule rule, which determines the priority of vehicles when passing through the intersection according to the travel time of individual vehicles. On this basis, we propose a new algorithm to achieve the optimal solution with low overhead. Finally, we build the simulation model and implement the algorithm for performance evaluation. Comprehensive simulation results demonstrate the superiority of the proposed algorithm. I. I NTRODUCTION Road intersections today are the sources of much traffic congestion and many accidents [1, 2]. According to Urban Mobility Report 2012 [1], in each year, congestion wastes urban Americans 5.5 billion hours on traveling and an extra of 2.9 billion gallons of fuel. Vehicular Cyber Physical Systems (VCPS) have been proposed in recent years to take advantage of the latest advances in sensing, computing, communications and controlling technologies to improve the safety, efficiency and sustainability of the road transportation system. Especially, VCPS integrates autonomous vehicles, vehicular communica- tions and controlling technologies into a whole system and make autonomous intersection control (AIC) system [3, 4] a promising option in the near future for not only improving This work was supported in part by the National Natural Science Foun- dation of China under Grant No. 61572088; Fundamental Research Funds for the Central Universities (Grant No. 106112015CDJZR185518); National 863 Program 2015AA015304; ICT R&D program of MSIP/IITP (14-824- 09-013, Resilient Cyber-Physical Systems Research) and GRL Program (2013K1A1A2A02078326) through NRF. (Corresponding author: Kai Liu.) road safety, but also enhancing transportation efficiency and travel experience [5–8]. Vehicle-to-infrastructure (V2I) com- munication enables the ICU to collect the traffic information and vehicle status in real-time. In the conceived AIC system, autonomous vehicles are commanded and coordinated by an ICU to cross the intersection cooperatively, because vehicles are able to accurately measure and control their kinematic states by equipped sensors. Autonomous intersection control is one of the most critical applications in VCPS. In recent years, cooperative intelli- gent driving techniques show great potentials in improving the performance of the intersection control [9–15]. However, previous research mainly focused on vehicle safety and traffic efficiency, while ignoring travel experience of passengers when designing the protocol. Collision avoidance is one of the most critical topics in AIC system. Hafner et.al [10] addressed the two-agent safety control problem for piecewise continuous system with disturbance and imperfect state infor- mation. Kowshik et.al [13] analyzed the vehicle behaviors and designed a hybrid architecture for guaranteeing the safety in the worst case behaviors. Besides collision avoidance, some researchers focused on improving traffic efficiency. Jin et.al [12] proposed a modified multi-agent system (MMAS) to provide system-wide benefit in traffic efficiency. Azimi et.al [9] designed a V2V intersection protocol called STIP to avoid vehicle collision and increase traffic throughput at intersection, which is suitable for both intersection and roundabout. Some studies [11, 14] designed the AIC system by transforming the problem into different forms of non-linear optimization models based on different purposes. These techniques require high computation overhead and thus they cannot achieve optimal solutions in real time. However, none of these work has con- sidered the issue of travel experience. For example, aggressive acceleration/deceleration actions may cause passenger sick or uncomfortable [16]. As far as we know, the closest AIC system to our work 2016 Intl IEEE Conferences on Ubiquitous Intelligence & Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People, and Smart World Congress 978-1-5090-2771-2/16 $31.00 © 2016 IEEE DOI 10.1109/UIC-ATC-ScalCom-CBDCom-IoP-SmartWorld.2016.49 203