Classification And Underwater Acoustic Communication Based On Dolphin Whistles

In recent years,underwater acoustic communication(UWAC)has been an important research area for the development of marine resources and military operation.After decades of precipitation and development,UWAC has reached a significant level of scientific research achievements and technology.Formerly,scholars have been looking for a system with high speed and long transmission distance in underwater acoustic channel.However,with the urgent need of national defense security,safety and covertness have become the main trend and direction of the development of UWAC.Previously covert underwater acoustic communication sustained the opinion of the low signal-to-noise ratio by reducing the transmitting power of the communication signal,making the communication signal be annihilated by marine environment noise.However,this technique has a severe disadvantage that the communication distance is limited,which making it impossible to realize long-distance UAC.Additionally the communication is weak in covertness,and the safety of the launching platform is seriously threatened in the sea.In order to ensure the security of UWAC system,a covert UWAC technique based on bionics is introduced.The covert UWAC at all distances could be achieved by mimicking the sea natural noise as the information carrier.The mimicked signal can be detected by the enemy but it is excluded from the process of recognition due to similarity with the sea natural sound,giving perfect low probability of recognition characteristics.It reduces the recognition probability in the communication process and promises the security of UWAC effectively.Based on the idea of bionic camouflage,the dolphin whistles are used as information carrier in this paper.Methods of signal processing and covert underwater acoustic communication based on dolphin whistles are studied.The work is shown as follows:1.The vocal mechanism of dolphins is analyzed from a biological perspective,including the sound field distribution and beam directivity,which is useful for understanding the principle and function of biological sonar.Four samples of dolphin whistles are selected to analyze the acoustic features and characteristics.2.The automatic extraction methods of whistle contours are studied.After signal processing,the extraction algorithm of single whistle contour based on particle filter and that of multi-whistle contours based on Gaussian-mixture probability hypothesis density filter are studied respectively.The feasibility of the two methods was verified by MATLAB.3.Classification and recognition methods based on whistle signals are studied.According to the characteristics of the dolphin whistles in time-frequency domains,whistle signals are divided into six categories.Based on the classification rules,two methods for automatic classification and recognition of whistles are studied: one is LBP and KNN joint algorithm,and the other is deep learning algorithm based on convolutional neural network.The effectiveness of the two recognition algorithms was verified by MATLAB.4.A novel method of covert UWAC based on dolphin whistles is researched.Dolphin whistles are used as the communication carriers.Combined with the excellent characteristics of the audio steganography algorithm in the transform domain,the whistles are transformed by DCT.The secret information is modulated in the DCT domain to enhance the covertness of the communication system.The m series is employed to extend the information and improve the reliability.The performance of the communication method was verified by simulation or pool experiments.In summary,this paper studies the methods of contour extraction,classification recognition and covert UWAC based on dolphin whistles.The feasibility and effectiveness of the algorithms are verified by simulation or pool experiment.