Research On Target Spectrum Line Autonomous Detection Technology

Due to its excellent flexibility and security,underwater unmanned systems can perform relatively dangerous tasks in the combat process,reduce operational casualties,and facilitate cluster operations.They can greatly expand the naval combat capabilities and have important military value.As the "eyes and ears" of underwater unmanned systems,sonar systems have an important position in underwater unmanned systems.Due to the advantages of stable energy concentration and long propagation distance,the line spectral components in the ship’s radiated noise can significantly improve the detection performance of passive sonars on low-noise targets.Aiming at the characteristics of unattended underwater systems,changing working environments,and changing motion states,this paper investigates the spectrum line autonomous detection algorithm and spectrum line autonomous tracking algorithm,and obtains different lines under different conditions.The advantages and disadvantages of the autonomous spectrum detection algorithm and different spectrum line autonomous tracking algorithms.This paper studies three power spectrum estimation algorithms,and compares three power spectrum estimation methods in different situations: the power spectrum estimation method based on the average periodogram method is a more robust power spectrum estimation method,which has Relatively stable detection effect;Power spectrum estimation method based on phase compensation has excellent detection performance when the online spectrum is stable,but the detection performance decreases when the frequency and phase of the online spectrum change;the power spectrum estimation method based on sparse reconstruction is It has good detection performance under the background of white noise,but because the sparsity is destroyed under the background of colored noise,the detection performance is reduced and the amount of calculation is large.In view of the changing working environment of the underwater unmanned platform,this paper also researches the background autonomous estimation algorithm,which can adaptively estimate the continuous spectrum background based on the power spectrum estimation results.A spectrum line autonomous extraction algorithm is presented.According to the stable existence of spectrum line in time,this paper studies three algorithms of spectrum line autonomous tracking,and compares the three algorithms in different cases.The spectrum line tracking algorithm based on edge detection has a simple algorithm and a small amount of calculation.Advantages,but under the condition of low signal-to-noise ratio,the ability to suppress noise is weak,and the ability to connect the discontinuous spectrum line is weakened;the spectrum line tracking algorithm based on Kohonen neural network can complete spectrum line tracking in various situations,but it will Isolated noise points are connected to form a false spectral line.The width of the spectrum line tracked under Doppler conditions is very wide and the amount of calculation is huge.The spectrum line tracking algorithm based on the hidden Markov model has the following features in the cases discussed in this chapter.Very good tracking effect,but it has a large amount of calculation,a slow operation speed,a large memory consumption,and needs to realize the predicted maximum number of line spectra.In this paper,the experimental data collected by the passive sonar on the unmanned platform are processed using various methods,and the processing results are compared and analyzed.Each algorithm can effectively observe the transmitted signals and verify the effectiveness of each algorithm.In order to verify the practicability of the algorithm,TMS320C6678 was used to design and implement the real-time autonomous line spectrum extraction software.The software was tested.The correctness and real-time performance of the software were verified by processing the measured data collected from the pool.