無線網路寬頻語音通訊服務的品質管理研究

Abstract

手持行動電話已是無線個人通訊必備的工具，網際網路的興起更成為資訊傳播的重要平台，結合兩者的整合式數位語音服務將展現出前所未有的應用需求與商機。本計畫將建構一無線網路寬頻語音傳輸架構，進一步探討網路服務品質的管理機制，以提高網路語音傳輸的通話品質與強健性。本計畫以三年為期，逐年探討三個相關課題：(1)強健性寬頻語音通訊系統的製作；(2)整合訊源與通道訊息的疊代解碼演算法；(3)多重串流封包的播放排程設計。 第一年度研究規劃主要著重於寬頻語音的適應性多重速率編碼系統製作，以期能因應通道環境的改變而調整其編碼模式與錯誤更正位元配置。其關鍵在於將線性預估係數執行分次多階向量量化的壓縮處理，但網路封包漏失會嚴重影響通話品質。因此我們將採用多重敘述向量量化的傳輸模式，並重新規劃其碼書索引指定設計成為一個最佳化計算的理想匹配問題，最後將參考基因法則建立其全域最佳解的隨機搜尋演算法。 第二年度主要是設計多重敘述向量量化的合併訊源通道解碼機制。解碼演算法將兼顧即時製作與準確性，前者強調疊代演算得以快速實現其依序更新，後者則要求交叉運用存在於相鄰敘述間與錯誤更正位元的相關重複訊息。至於索引序列殘餘冗息的運用，將利用一階馬可夫隨機過程模擬，配合迴旋碼的軟性輸出訊息來估算傳輸索引值的後驗機率。 第三年度則提出一個基於音質預測模型的適應性播放排程演算法。主要構想是針對多重串流語音封包的播放排程設計，先規劃成為一個音質損害最小化問題，再尋求整體延遲與封包漏失的最佳平衡點。

The increasing use of wideband speech (50-7000 Hz) for interactive audio applications has lead to the adaptive multi-rate wideband (AMR-WB) speech coding algorithm standardized by ETSI/3GPP and ITU-T. However, packet loss and network delay are two essential problems to real-time wideband speech communication over the IP and mobile networks. The purpose of this three-year project is to develop a mobile wideband speech communication system and its quality of service (QoS) control for increased channel robustness. The first part of this project will focus on the error concealment of packet loss as well as channel bit errors. The basic strategy is a multiple description vector quantization (MDVQ) system, in which multiple correlated indexes of the source are assigned and transmitted over channels having largely uncorrelated loss and delay behavior. Finding the best index assignment is an NP-hard problem and hence we will develop a genetic-based stochastic search algorithm on the basis of a linear programming framework. The main attraction of genetic algorithm is that the given search space is explored in parallel by means of iterative modifications of a population of chromosomes. We will also introduce the use of turbo principle to develop an iterative source-channel decoding algorithm for better decoding of multiple descriptions transmitted over a noisy channel. Another important issue to address is the playout buffer design which is often used at the receiver to smooth out the jitter for timely reconstruction of the speech. We will formulate adaptive playout scheduling of multiple voice streams as a constrained optimization problem that leads to a better balance between end-to-end delay and packet loss. Also proposed is a perceptually motivated optimization criterion and a practically feasible algorithm for the playout buffer design.