Dynamic Throughput Estimation for Wireless Video Communications Wuttipong Kumwilaisak †, JongWon Kim ‡, Robert Ku †, and C.-C. Jay Kuo † † Integrated Media Systems Center and Department of Electrical Engineering University of Southern California, Los Angeles, CA 90089-2564 ‡ Department of Information and Communications Kwang-Ju Institure of Science and Technology, Kwang-Ju, Korea E-mail: kumwilai@scf.usc.edu, jongwon@kjist.ac.kr, cckuo@sipi.usc.edu ABSTRACT Dynamic throughput estimation for wireless multimedia transmission is examined in this research. A novel dynamic mapping scheme from measurements of a real-world wireless environment to a discrete Markov model is first proposed to simulate wireless video packet-based transmission. Then, the available throughput is estimated by using the derived discrete Markov model together with the number of negative acknowledgements (NACK) that are fed back from the receiver to the transmitter. The estimated throughput is used to perform the adaptive region-of-interest (ROI) rate control based on the multiple constraint optimization. The performance of the channel mapping scheme, the throughput estimation, and the region-of-interest rate control is demonstrated by simulating a micro-cell environment and ITU-T H.263+ video standard. Performance variation due to different estimation of throughput is compared and the performance improvement due to the multiple constraint optimization is demonstrated both in objective and subjective quality. 1. INTRODUCTION In the upcoming third-generation (3G) wireless system, packet-based transmission is the trend in multimedia services in both video conferencing and video streaming applications. Understanding the effect of the time-varying wireless channel, including the path loss, the long-term fading, and the short-term fading on the transmission packet, is one of the crucial tasks. The channel model is utilized to design and study the performance of the transmission protocol such as ARQ or FEC [1]. The Markov model is widely used to simulate the influence of a wireless channel to the transmitted packet in various research areas such as rate control [2] or joint source-channel rate allocation [3]. Most work in modelling the Markov model [1,4] has been concentrated on the short-term fading effect in packet-based transmission with the high data rate and the small packet size assumptions. However, with the operating data rate 64 ∼ 384 kbps and the transmitted packet size 100 ∼ 200 bytes in the 3G system [14], these assumptions may not be valid. Based on this fact, the Markov model with the long-term fading effect provides a more suitable solution to simulate the packet-based transmission system. To best represent the wireless channel, the parameters of the Markov model such as the transition probability can be computed from the measurement data [2] or from the wireless channel model [4]. However, the modelling Markov parameter based on the wireless channel model seems to have more advantageous than the direct measurement due to the reduced computational complexity and effort in collecting data. In this research, we propose a novel solution in constructing the Markov model based on both the long-term and the short-term fading effects to best represent the physical wireless channel for wireless multimedia transmission. In particular, we develop a mapping scheme that converts physical measurements of a real-world wireless environment to a discrete Markov model via an iterative algorithm. With such a mathematical channel model, we propose a simple yet efficient way to estimate the throughput, which employs both the number of negative acknowledgements (NACK) and the maximum likelihood principle. Next, the throughput estimation from the Markov model is used to simulate the region-of-interest (ROI) rate control for video transmission over the wireless channel. Adaptive rate allocation with multiple constraints, including the estimated throughput and the region quality requirement, is used to control the quality of each video region. The region in a video frame can be obtained from either user intervention or automatic machine segmentation. Finally, a packetized transmission scenario is set up by the ITU-T H.324M wireless video framework. In the experiment, error-resilient H.263+ video is multiplexed packet by packet with a