| The on-the-move communication systems have been widely used. The usage of conventional on-the-move communication systems with mechanical scanning antennas has been limited, because of the high price and poor tracking accuracy. At present, the systems with phased array antenna are on sale. However, the price is still too high. Thus, how to reduce the cost of the system while maintain the tracking accuracy is becoming the key of the research.We research deeply in the beamforming of the on-the-move communication system with phased array antenna. In this thesis, we propose the simultaneous perturbation stochastic gradient estimation algorithm based phase shifter control voltage beamforming technology, and set forth the beam space based and singular value decomposition based beamforming technologies, which do not require the knowledge of signal’s DOA (direction of angle, DOA). Thus we named these three beamforming technologies as unknown DOA algorithm. For the high cost and bad tracking accuracy of the beam space based beamforming technology, we take the phase shifter control voltage based and singular value decomposition based beamforming technologies in-depth analysis only.Phase shifter control voltage based beamforming technology uses analog phase shifters, and beamforming in the microwave stage, hence, it requires only one receiver. The aim of the algorithm is maximizing the receiving power by adjusting the control voltage of the phase shifters. We solve the optimization function by using the deepest descent algorithm. In this thesis, the simultaneous perturbation stochastic gradient estimation algorithm is used to solve the gradient of power w.r.t. voltage. Compares with the other approximation algorithms, the simultaneous perturbation stochastic gradient estimation algorithm can evaluate all the elements of the gradient by one-time, no matter how much the dimension of the gradient is. Thus the simultaneous perturbation stochastic gradient estimation algorithm has a high convergence speed. For the purpose of fast convergence, we discuss the value of the disturbance and the step-size deeply, and let the step-size changes with the iterations and the space of the array, which changes with the iterations only in the prototype algorithm, and point out the parameter selection criteria when calculate the disturbance. Finally, the impact of the noise to the algorithm has also been discussed.The singular value decomposition based beamforming technology is used for On-the-Move Communication. The algorithm is a digital beamforming (DBF) algorithm, thus processes the received data directly, and compensate the phase difference of the signal reaches the different element of the array when processing the data. Thus, the phase shifters and other hardwares are not required. For the reason of process the receive data directly, the singular value decomposition based beamforming can implement real-time in the tracking moment. And comparing with the phase shifter control voltage based beamforming technology, the advantage is obvious. Since the singular value decomposition based beamforming technology needs more receiving channels compared to the phase shifter control voltage based beamforming technology, and for the aim of reducing the costs, we should partition the array to several sub-arrays, hence, only the number of the sub-arrays receiving channels is needed. Based on the sidelobe level and mainlobe deviation, a method to calculate the number of the subarrays is studied. Then, the weight of each element in every sub-array is deeply studied. We construct a model which gets the lowest maximum sidelobe level in the scanning field. Finally, the impact of the noise to the algorithm is also been discussed, noting that the algorithm only adapt to the high SNR environment.Finally, we compare the two beamforming techniques, and find each has its advantages. The MATLAB simulations show that both of the two algorithms we have discussed above can track the change of the signal’s DOA effectively. |
PDF Full Text Request