TABLE OF CONTENTS As discussed earlier, tubo codes have an amazing error correcting capability and are, therefore, very attractive for many applications. In this chapter, we address some of the implementation issues. The complexity of a turbo-decoder is much higher than the complexity of the encoder. Thus, we put emphasis on the decoder. Low cost and low power consumption are extremely important issues for turbo decoder implementation. Consequently, fixed-point arithmetic and fixed-point implementation are unavoidable issues. The Max-Log-MAP algorithm discussed in this book is simple enough and performs very close to the MAP algorithm. So, it is a good trade-off between the complexity and performance. In this chapter, the effect of input data quantization for TCC and BTC are presented. Moreover, the effect of correction term in Max-log-MAP algorithm is discussed.
So far, we have assumed perfect carrier phase and channel SNR estimation in our simulations. However, this does not truly represent a practical system where channel impairments caused by the noise and attenuation occur. Some examples of channel impairments are carrier phase offset and channel SNR mismatch, i.e., a difference between the assumed and actual values of the phase and the SNR. In this chapter, the effect of channel impairements on turbo codes are presented. We will also discuss hardware implementation for turbo codes on FPGA, ASIC and DSP.
Algorithm used in a turbo decoder are usually specified in the floating point domain. Fixed point number representation is mandatory for most target architectures, thus...
