A noise-tolerant system for protocol formats extraction from binary data

Under the wireless communication network environment, the received signal will be effected by noise and interference. And it will produce many inaccurate codes consequentially. Under this situation, the error bits in the binary data after demodulating and decoding will affect the formats extraction of the unknown protocols. In order to improve the accuracy under the noise interference and the adaptability of binary data type, this paper proposed a novel multiple sequence alignment algorithm based on the approximate matching (MSAAM) to process the data with random error bits. Bad sequence erasing was introduced to reduce the interference from the burst error bits. After the sequence alignment, a noise-tolerant system for protocol formats extraction from binary data was established combining statistical analysis and characteristic library. The experiments showed that the system was robust to error bits and can infer part of the format information and the functions of some fields.