Uncoded symbol error rate estimation: methods and analysis

In wireless communications systems, channel quality estimates are often used to provide a measure of the quality of service or to enable resource allocation techniques that improve system capacity and/or user quality. The uncoded bit or symbol error rate (SER) is specified as a channel quality metric in second and third generation cellular systems (e.g., general packet radio service (GPRS), enhanced general packet radio service (EGPRS), and universal mobile telecommunications system (UMTS). Nonetheless, techniques to estimate the uncoded SER are typically outside the scope of these wireless standards and are not specified. In this paper, we analyze the performance of a number of uncoded SER estimation techniques, including a novel technique in which we use the soft information of the received symbols to obtain a fast and accurate estimate of the uncoded SER. The technique we introduce has been found to outperform, in terms of accuracy and required estimation interval, conventional hard-decision based techniques that use test-patterns, or use a decode/re-encode/compare approach . Our technique also outperforms the brute-force technique, which is to send a known test-pattern, demodulate it at the receiver, and count the observed discrepancies.