Hydrophone Ultra Complex Model And Its Doa Estimation Algorithm

In the field of acoustic signal processing, classic hydrophone can't utilize the spatial information in a single sensor mode. So it is required to deploy several sensors in some geometrical form to determine the DOA (DOA Direction of Arrival). Recently, a novel acoustic sensor has attracted widely interests, namely the acoustic vector sensor. Acoustic vector sensor has four output signals which correspond to the pressure and the three velocity components along three orthogonal axes. Acoustic vector sensor has many advantages compared with scalar acoustic sensor. However traditional technology of processing the vector sensor signals is to concatenate the four components, then process the signals with methods using in the processing of hydrophone signals. Because these traditional signal processing methods violate the relationship of the four components themselves, these methods have some disadvantages.The form of acoustic vector sensor signal is similar to the hypercomplex algebra which was proposed by Hamilton in 1843. Hypercomplex deals the four components holistic in the process of operation, so the relationship of every component can be maintained. Due to this fact, representing the outputs of acoustic vector sensor signals will bring many advantages. Based on the structure of Hypercomplex algebra and output signals of vector acoustic sensor, this paper proposes the hypercomplex measurement equation of the vector acoustic sensor and the hypercomplex measurement is extended to the acoustic vector sensor array.Based on the hypercomplex measurement equation, a novel DOA estimation algorithm is proposed. The concepts of energy and energy distribution under hypercomplex algebra are defined in this paper. The DOA algorithm is proposed through studying the energy distribution over the hypercomplex space. The proposed DOA algorithm has explicit physical meaning that the energy reaches its maximum value along its direction of propagation. Furthermore the relationship of acoustic vector signal energy under hypercomplex algebra and the base of simplicity decomposition is given. Finally the relationship between estimated value and the hypercomplex decomposition axis is also given.Based on our hypercomplex model, the hypercomplex MUSIC-like algorithm can be used to estimate the directions of arrival of the sound sources. The analytical results show that, using some signal processing algorithm with the proposed hypercomplex model of the acoustic vector sensor, the memory requirements for representing the data covariance model are reduced 75% while the division requirements are reduced by 75%, for equivalent performance when compared with the long vector method. The results show that the hypercomplex model of acoustic vector sensor breaks a new path for the signal processing field.