PERCEPTUALLY-OPTIMIZED PACKET SCHEDULING FOR REAL-TIME 802.11 VIDEO COMMUNICATIONS BETWEEN VEHICLES Enrico Masala, Juan Carlos De Martin Computer and Control Engineering Dept., Politecnico di Torino, I-10129 Torino, Italy {masala,demartin}@polito.it ABSTRACT The automotive industry is increasingly looking at solutions for intervehicle wireless communications for applications ranging from safety to entertainment. Warning signals are the most immediate applications, but more complex forms of communications, and video in particular, could be used by innovative applications such as multi-vehicle-based visual processing of road information, multi-vehicle radar systems for obstacle avoidance and automatic driving, and more generally swarm communications among cars traveling along the same road. In this paper we present a perceptually-optimized packet scheduling algorithm which can transmit video data captured by an on-board camera to another vehicle in proximity, using standard 802.11 wireless technology. Low-delay transmission of on-board captured H.264 video is simulated using actual intervehicle packet transmission traces. The results show that the proposed perceptually-optimized algorithm achieves a consistently higher quality compared to two reference techniques, with gains up to 2 dB PSNR. Moreover, timely dropping of packets which cannot reach the receiver on time for playback has been shown to greatly improve the performance of the transmission system. Index Terms— Packet scheduling optimization, low- delay video, intervehicle communications 1. INTRODUCTION The potential applications of intervehicle wireless communications are numerous, ranging from safety to entertainment. Warning signals are the most immediate applications, but more complex forms of communications, and video in particular, could be used by innovative applications such as multi-vehicle-based visual processing of road information for obstacle avoidance and automatic driving, and more generally swarm communications among cars traveling along the same road. Numerous issues must be addressed for efficient intervehicle communication. The Wireless Access in Vehicular Environment (WAVE) suite of communications standards, that has recently been approved for trial use, aim at addressing issues such as security, management of multiple radio channels and system resources. Other standards are under active development, such as the IEEE 802.11p, an extension of the well-known 802.11, which is expected to cover protocol and networking services in WAVE. Intervehicle and vehicle to infrastructure communications, however, have already been experimented by several research projects based, for instance, using the 802.11 standard [1][2]. Those works mainly focused on throughput and connectivity issues. Others investigated routing [3] and information dissemination issues [4]. Modifications of the existing MAC access protocols have also been proposed [5] to provide important features such as bounded maximum access delay. However, for more specific applications, such as real-time multimedia as opposed to generic data communication, it is possible to exploit the peculiar characteristics of the data, and in particular their non-uniform importance, to further optimize the communication performance. In this work we present a perceptually-optimized packet scheduling algorithm aimed at providing reliable low-delay intervehicle video communications. The proposed algorithm is tested using actual intervehicle packet transmission traces and video sequences captured from on-board cameras, and results are compared with two reference techniques, that is the standard 802.11 MAC-layer retransmission scheme and a delay-constrained retransmission scheme. The paper is organized as follows. Section 2 briefly introduces the H.264 standard, explains the distortion estimation technique and discusses the main issues of multimedia transmission over 802.11. The proposed perceptually-optimized scheduling technique as well as the two reference techniques are analyzed in details in Section 3. Section 4 describes the experimental setup and Section 5 discusses the results. Finally, conclusions are drawn in Section 6. 2. H.264 VIDEO COMMUNICATIONS OVER 802.11 2.1. The H.264 Video Coding Standard In this work we consider video communications based on