TrCH BER Estimation for Turbo Code
Original Publication Date: 2003-Jan-21
Included in the Prior Art Database: 2003-Jan-21
AbstractThis paper presents a simple method of TrCH BER estimation for Turbo Codes as required in 3GPP base station. The TrCH BER is estimated by counting differences between systematic bits pre-and post-decoding. Consequently there is no need for re-encoding.
TrCH BER Estimation for Turbo Code [MP1]
This paper presents a simple method of TrCH BER estimation for Turbo Codes as required in 3GPP base station. The TrCH BER is estimated by counting differences between systematic bits pre- and post-decoding. Consequently there is no need for re-encoding.
It is required in 3GPP that a quality estimate of transport channel (TrCH) in terms of bit error rate (BER) be reported to higher layer. The measurement is defined as the error rate measured for one TrCH during a single transmission time interval (TTI) at the decoder input excluding punctured bits. The average of this measurement should be within 10% of the true TrCH BER under the conditions that the averaging over successive TTIs such that at least 500 errors occur and the actual TrCH BER is less than 15%.
To meet this accuracy requirement, it is widely understood that the TrCH BER is estimated by re-encoding, that is, the decoded information bits for a TTI are passed through the corresponding encoder, puncturing/repetition procedure and channel symbol mapping then compared with sliced decoder input. However, this re-encoding not only introduces extra complexity but also additional latency. To avoid this, another scheme, which is based on estimated systematic mean and variance and under AWGN signal model assumption was designed. Unfortunately, it has been found that this method can not meet 3GPP accuracy requirement. Clearly, it is highly desired that a simple BER estimation with required accuracy be developed. In this invention, a very simple method with accuracy very close to re-encoding is presented. Since this invented method does not need re-encoding, we can save complexity associated with encoding, such as encoder and puncturing/repetition operation, and eliminate latency corresponding to re-encoding.
In the following, we give a brief review on turbo code. Consider a typical turbo encoder that is a parallel concatenation of two recursive systematic convolutional (RSC) encoders with an interleaver between them as illustrated in Fig. 1. The output of the turbo encoder is generated by puncturing (for different coding rate) and multiplexing the information bits , parity bits , and tail bits, where N denotes the number of information bits in one turbo code block.
The output of turbo encoder is depicted in Figure 2. The tail bits are used to force turbo encoder back to all-zero state at the end of code block. For 3GPP specified turbo code, number of tail bits is 12.
At receiver, turbo decoding is achieved in an iterative fashion as shown in Fig. 3. The turbo decoder basically consists of two constituent decoders, denoted DEC1 and DEC2. The received signal is de-multiplexed intoxk, y1kand y2k representing samples for systematic bit bk, parity bits p1k and p2k respectively, as shown in Fig. 4. DEC1 and DEC2 are soft-in-soft-out (SISO) recursive systematic convoluti...