CS-CPA: Packet Loss Recovery for Audio Multimedia Streaming Based on Chained Compressed Sensing

Abstract

Packet loss is an inevitable problem confronted with real-time audio transmission over IP network. Recently, a packet loss recovery scheme for audio multimedia streaming based on compressed sensing (PLRCS) has been proposed. While PLRCS is a promising scheme to resist the packet loss problem, it has poor recovery performance when the packet loss rate (PLR) is high. On the other hand, to improve the data security and reduce the data volume, a chained CS (CCS) has also been proposed to provide lightweight encryption of data. However, the CCS cannot be applied to the packet loss problem directly. In this paper, combining CS and chain technology, we propose a packet loss recovery scheme, namely CS-CPA, which is equivalent to designing a structurally random matrix (SRM) by incorporating a predetermined spare binary matrix (SBM) and the packet loss process in the IP network. Experimental and analytical results show that, the proposed CS-CPA is not only superior to previous methods in terms of the recoverability, but also can resist several potential attacks, such as ciphertext-only attacks (COA) and known-plaintext attacks (KPA).