基于RTP协议的流媒体数据容错识别算法研究

Research on Fault Tolerant Recognition Algorithm of Streaming Media Data Based on RTP Protocol

针对流媒体数据在无线传输过程中因传输差错造成受损数据包无法识别利用的问题,提出一种基于关键协议字段修复的RTP数据容错识别算法。通过对承载RTP报文的IP数据包各协议字段的分析,确定对RTP数据识别有重要作用的关键协议字段;通过对正确RTP数据的统计分析,获取关键协议字段的先验信息,最后依据最大后验概率准则对含错关键协议字段进行修复。实验结果表明,与现有算法相比,该算法对RTP数据流的召回率明显提升,网络数据的利用率显著提高。

To identity and utilize corrupted packets caused by transmission error in wireless environ- ment, a RTP data streaming fault tolerant-recognition algorithm is proposed based on critical field. Critical fields vital to RTP data stream recognizing is determined by analyzing IP fields; priori intor- mation of critical fields is acquired by statistically analyzing correct RTP packets. Corrupted critical fields are repaired according to maximum a posterior（MAP）. Experiment results show that the algo- rithm could obtain higher RTP stream recall rate than traditional method, and enhance utilization of network data obviously.