On the Error Performance of Coding and Equalization in Low-Complexity Ultra-Wideband Communication Systems

In this paper, the performance of various channel coding schemes is investigated in pulse-based ultra-wideband (UWB) communication systems for applications in short-range indoor environments. Pulse-based binary (BPSK) modulation and decision-feedback equalization (DFE) is considered. Con- catenated adaptive equalization and coding is explored as an alternative to the more complex and often impractical joint coding and equalization. A block length of approximately 1000 bits is considered in this paper as it results in a static channel with minimal latency while still yielding relatively good error performance. The error performance of a previously proposed turbo product code (TPC), based on two identical Hamming (31,26) codes, is simulated and compared with that of other channel coding schemes of similar rate and code length. These include a regular LDPC (1057,813) code, a memory-6 rate-3/4 punctured convolutional code, a Reed-Solomon (127,89) code and a concatenated (off-the-shelf) code with a Reed-Solomon (255,239) outer code and a memory-6 rate-3/4 punctured con- volutional inner code. The inclusion of the concatenated Reed- Solomon scheme serves as a reference, as this is an off-the-shelf classical and still popular solution. The simulation results show that, among the coding schemes considered, the LDPC code offers the best error performance. personal-area network (WPAN) communication system (6). These requirements include a bit rate of at least 110 Mbps at a distance of up to 10 m and 200 Mbps at up to 4 m, with desired rates up to 480 Mbps. A pulse-based UWB approach is a good candidate for meeting the WPAN requirements. In this paper, the error performance of error correcting coding is investigated in a UWB communication system with adaptive decision-feedback equalization and pulse-based mod- ulation for applications in short-range indoor environments. Performance is measured using the reference UWB multipath channel models generated by the IEEE 802.15.3a standard group (7). A communication system is considered here in which known symbols are used to train the adaptive equalizer. To reduce the variation in the bit error rate with respect to changing multipath channel conditions, a turbo product code (TPC) with two component (31, 26, 3) Hamming codes was proposed in (9). The purpose of the work reported here is to consider other channel coding schemes of similar rate, such as a low-complexity LDPC code, a Reed-Solomon code and a concatenated code, and to evaluate and compare their error performance with respect to the TPC code previously proposed. The rest of the paper is organized as follows: In section II, the UWB communication system model used in this study is presented. It is shown that an equivalent symbol-spaced UWB channel model is obtained between the input to the pulse shaping filter and the output of the matched filter. Section III considers the performance of suboptimal adaptive equalizers as low-complexity alternatives to the Viterbi equalizer. Perfor- mance is studied in terms of equalizer length and sensitivity to different channel realizations. Simulation results, presented in section IV-A, of combinations of adaptive equalizers and a turbo product code, illustrate the benefits of using channel coding as an effective way to improve performance and reduce sensitivity to channel variations.