Research On The Algorithm Of Telemetry And Remote Control Signal Analysis And GPU Implementation Technology

With the development of modern technology,Telemetry and remote control technologies are increasingly used in military and civilian applications.The analysis and identification technology of telemetry and remote control signals has also become a key part of it.At the same time,with the increase in the amount of signal data and the increase in the complexity of the analysis algorithms,the traditional signal detection and recognition algorithms based on serial computing by the CPU have been unable to meet the real-time requirements of signal processing.The rapid development of GPU general-purpose computing technology represented by CUDA has provided a technical approach for the rapid analysis and identification of signals.This thesis focuses on the analysis and identification of telemetry and remote control signals and their GPU implementation.The main work is as follows:1.This thesis introduces the CUDA programming model,execution model and memory model,and optimization strategy of CUDA program.2.Aiming at the problem of random access of DSSS signals,due to the relatively low signal noise and Doppler frequency shift,the detection of DSSS signals becomes more difficult and the real-time performance is poor.In this thesis,a DSSS signal detection system based on GPU is designed.It can use less a priori information to complete the signal detection and parameters estimation of DSSS signals with Doppler frequency shift in a low SNR condition.This article first introduces the specific algorithm principle and implementation process of each module of the DSSS signal detection system.Based on this,the parallelism of each module of the system is studied,and the parallel algorithm design of each module under the CPU + GPU architecture is completed.The simulation data verifies the correctness of the GPU implementation of each module,statistics the GPU acceleration effect of each module,and conducts performance analysis.The results show that compared with serial programs under MATAB,the GPU implementation of each module can achieve a speedup of more than 2 times,and the GPU implementation of the symbol width estimation module can achieve an acceleration ratio of 13.201 times after optimization.3.Aiming at the problem of random access of six types of digital transmission signals: BPSK,QPSK,OQPSK,UQPSK,GMSK,PCM / FM,this thesis designs a GPU-based digital signal detection system,which can be used to quickly realize the signal detection,modulation type identification and parameter estimation of digital transmission signals with large code rate range containing Doppler frequency shift.This thesis first designs a CPU-based digital signal detection system,and uses simulation data to verify its performance in a MATALB environment.Then,based on the implementation process of each module of the system,the parallelism of each module is analyzed,a parallel acceleration strategy is proposed for each module,the GPU implementation of the system is completed,and the performance of the signal detection module under GPU implementation is analyzed using simulation data Acceleration ratio.The results show that the signal detection module realized by GPU can speed up more than 2 times than CPU.4.This article combines the DSSS signal detection system and the digital signal detection system into a unified telemetry remote signal analysis system,which can simultaneously complete the signal detection,modulation type identification and parameter estimation of DSSS signal with a signal-to-noise ratio of-18 dB and the digital signal with a signal-to-noise ratio of 5dB or more.After that,this thesis verifies the performance of the combined system running on the GPU and gives the results of the speedup.The results show that the acceleration ratio of the system implemented under GPU will vary with the change of signal parameters,and the acceleration ratio of the system is 2 times or more in different situations.