Research On Beamspace-based DOA Estimation For Ultra Wideband Signals

Ultra-wideband (UWB) technology has broad application in wirelesscommunication, radar, tracking, imaging and many other fields. Recent years, withthe development of UWB technology, it become possible to implenmatation arrayantenna in UWB system, and the direction of arrival (DOA) estimation of UWBsignals becomes one of the key technologies of UWB communication system andan important research direction for current personal communication, particularlyfor short-range wireless communication. Now DOA estimation of UWB signals hasjust started, this paper gives deeply research on the problem of DOA estimation forUWB signals based on the broadband CSM algorithm and beam-space algorithm.First, in this paper, DOA estimation method of UWB signals based on CSMalgorithm is studied. This method can be used to estimate coherent UWB sources,however, when the coherent sources are closed to each other, it can't estimate well.In view of this shortcoming, the paper suggested Forward/backward SpatialSmoothing (FBSS) method to improve CSM-UWB DOA estimation algorithm, theimproved algorithm can estimate space adjacent UWB sources well. We use thealgorithm in different channels, simulation results indicate that the algorithm in thechannel CM1 and CM3 are both able to estimate the corresponding UWB signalangles, and the result in line-of-sight environment is superior to the results in nonline-of-sight environment. As two-dimensional (2D) DOA can estimate azimuthangle and elevation angle at the same time, we propose a 2D DOA algorithm ofUWB signal based on CSM. The simulation results prove the correctness of thealgorithm.Secondly, because DOA estimation method in beamspace can utilize the prioriinformation of signals as well as the high Robust of beamforming method, thecomputation load and sensitivity to error of DOA estimation can be reduced at thesame time by beamforming process. In specific conditions, the methods inbeamspace show better statistical characteristics and practical value than methodsin element space. After collecting and summing up a lot of literature about UWBcharacteristic as well as beamspace-based DOA algorithms, we carefully study andanalyze the feasibility and advantages of beamspace-based DOA for UWB signals,and derive Chebyshev weight matrix and Discrete Fourier Transform weight matrix.Then we propose two DOA estimation methods for UWB signals based on frequency-domain frequency-invariant beamformers (FDFIB) algorithm andbeamspace high-Resolution (BSHR) method under different weights, respectively.Simulation results show that the two proposed algorithms can estimate DOAsof UWB signals.The comparison of different algorithms under different weightsshow that, under the same weight, the DOA estimation method of UWB signalsbased on BSHR algorithm has higher resolution capability, especially in the higherSNR circumstance, BSHR-UWB DOA estimation method has better estimationperformance. Comparison of different weights using the same algorithm, we foundthat in lower SNR circumstances, the DOA estimation method of UWB signalsunder DFT weight has smaller mean square error as well as better identifyperformance. Overall, BSHR-UWB DOA algorithm under DFT weights has betterestimation performance than other method. Simulation results with differentparameters under the DFT-BSHR DOA algorithm show that greater number ofantenna array, narrower beam width and more number beams, the algorithm willhave smaller mean-square error and better identify performance. Comparingbetween element space algorithm and beamspace algorithm, we found in thesmaller SNR circumstances, when UWB-DOA algorithm in elements space isalmost impossible to estimate the DOAs, beamspace algorithm can still estimateDOAs which prove that beamspace can reduce the SNR threshold. We use theDFT-BSHR DOA algorithm in different channels, simulation results indicate thatthe algorithm in the channel CM1 and CM3 both able to estimate the correspondingUWB signal angles, and the result in line-of-sight environment is superior to theresults in non-line-of-sight environment.Finally, the problem of beamspace-based DOA method for UWB signals underuniform circular array (UCA) is studied. Because the work above is all based onuniform linear array (ULA), and ULA has its own inherent weaknesses, such as itonly can give degree unfaintness angles, and from the Pattern we know tha180? tResolution is highest in normal line direction and the worst in the axial direction,so truly effective scope is only 120? degree. In response to this shortcoming, westudy characters of uniform circular array UCA and Mode-space transformingcarefully, propose beamspace DOA method for UWB signals under UCA. From thesimulation results show that the proposed algorithm can give 360? unfaintnessestimation, and obtain high resolution estimation.