Research On Channel Coding And Decoding In A Mutli-core DSP

Nowaday,channel coding has been widely used in communication systems to meet the higher and higher requirement of communication quality.This dissertation investigates two channel coding schemes,i.e.,convolutional code and Turbo code,and implements them in a multi-core DSP.First,this dissertation presents an overview of channel coding technologies,including the research and applications of convolutional code and Turbo code.In this dissertation,convolutional code and Turbo code are applied for the signaling data and traffic data of an aeronautical communication system respectly.Second,this dissertation introduces the structure of the the aeronautical communication link as well as the hardware platform.The TMS320C667 x DSP series,the SYS/BIOS multi-core development structure,the Turbo decoder coprocessor TCP3 D and the SRIO interface are described.Third,this dissertation describes the structure of convolutional code and implements the convolutional encoding and decoding based on C language.The influence of some specific parameters,such as the constraint length,code rate and the decision mode,on the code performance is investigated by simulation.Based on the simulation results,the convolutional code with a constraint length of 7 and code rate of 1/2 is selected.Decision mode will be adjusted according to the different modulation methods.Moreover,the performance of simple block parallel decoding and overlapping block parallel decoding is compared by simulation.Based on the results,the four-block overlapping parallel decoding algorithm is used as the decoding method.Then the convolutional code encoding and decoding are implemented in a multi-core DSP.Fourth,this dissertation introduces the structure of Turbo code and implements Turbo encoding and decoding based on C language.By simulation,a recursive systematic convolutional code with the constraint length of 4 is selected.The 3GPP interleaver generator is selected.The code length is set to be 3136 bits and the code rate is set to be 1/2.MAX-Log-MAP algorithm is used as the decoding method,and the number of iterations is set to be 4.Several block parallel decoding algorithms,such as the simple block parallel decoding,the overlapping block parallel decoding and the boundary condition iterative block decoding,are compared to reduce the decoding delay.Then the chosen Turbo code encoding algorithms are implemented in a multi-core DSP.Finally,the performance of convolutional code and Turbo code implemented in a multi-core DSP are tested.The results show that the DSP based implementation can achieve nearly the same performance as that of simulation.Meanwhile,the consumed time due to coding and decoding is tested to validate that the processing time can meet the system requirement.Furthermore,the performance of the whole link is tested in different aeronautical communication scenarios to verify the effectiveness of the implementation.