The Research Of Turbo Decoding Algorithm

From the thesis "the theories of mathematics in communications" announced by Shannon in 1948, the coding in channel has obtained very great achievement. Turbo codes were put forward by French scholar C.Berrou etc in 1993. Turbo codes have good capability near to the extreme limit in Shannon. It not only has eximious performance in the circumstance in low SNR, but also has a strong ability to keep from decline and interference, so turbo codes became the hotspot in coding domain and the breakthrough in the scopes of correcting right codes. The finding of the Turbo codes is the breakthrough in the scopes of correcting right codes in recent years. The emergence of the Turbo codes is gone by the name of milestone in the scopes of correcting right codes. As the theoretical principle becomes perfectly day by day, Turbo code is used to applied domain. Now it has already been extensively used in many communication protocols such as CDMA2000, the WCDMA, and the 3 GPP etc.The operation in Turbo codes is very complicated, and it takes up physical memory space bigger, and so it is difficulty in engineering achieving. This thesis introduced the development of theories and technique in signal channel coding first and intensive studied the basal theories on coding and encoding arithmetic of turbo codes. Then it brief explained the basic theories in coding and interweaver and encoding and status quo. In the basic of studying the arithmetic in encoding, this thesis primarily studies the arithmetic "MAP" and the arithmetic "Log-MAP" and the arithmetic "MAX-Log-MAP". According to the experiment, the arithmetic "MAP" is near to the arithmetic "Log-MAP" .And because the arithmetic "MAX-Log-MAP" has many plus in information, it is much loss in capability. But it is easy to hardware experimental results.Finally, investigating the interrelated problem which achieved encoder of turbo codes using fixed point DSP chip. Discussing many problems about the precision of quantification of coding,the disposal of overflow and deposited variable and so on. And draw a conclusion about the method of avoided overflow, which effectively avoid the overflow in the process of operation. The thesis uses C language to validate the arithmetic. And in the DSP chip TMS320C5416 in the company TI. Its capability is near to the coding in floating point numbers by testing. According to the experiment, the turbo encoder has a good capability and it works very steady.