【作者】 陈明明；

【导师】 刘昊； 李海松；

【作者基本信息】 东南大学 ， 集成电路工程（专业学位）， 2022， 硕士

【摘要】 近年来,随着物联网与人工智能的快速发展,定位技术也获得了迅速的发展。与此同时,市场对定位技术的需求也逐渐提高。然而基于低功耗蓝牙(BLE)的AoA(Angle of Arrival)或AoD(Angle of Depature)的新的室内定位技术因多径干扰等问题影响了定位精度。对此,本论文实现了一种面向BLE室内定位测向的闭式解解相干抗多径算法,来解决影响BLE定位的多径干扰问题。本文根据BLE信号传输模型和阵列测向理论,首先对BLE AoA/AoD测向模型进行分析,发现存在多径干扰导致测向角度偏差较大且不稳定。因此,本文简化室内定位模型,简化为主要有一个直射径和强反射干扰径的模型,并提出一种四天线线性阵列闭式解的解相干算法。该算法主要原理是将第二、三和四天线接收到的数据与第一天线共轭相乘后归一化,简化成三维线性方程组,该方程组有唯一闭式解,实际数学意义是对入射信号协方差矩阵的第一列求解。此外,根据一般多径反射路径信号幅度小于直射路径信号的原理,通过判断相干信号幅度的大小可以区分闭式解中直射信号入射角和多径反射信号入射角。最后本文在验证完在两条多径入射情况下算法实现的效果后,进一步在多条多径条件下测试该算法解相干测向的能力,并进行算法复杂度分析,验证了该算法具有稳定测向且复杂度较低的优势。本文经过仿真验证,证明该算法对四天线阵列的BLE AoA/AoD估计具有可分辨多径、解一定收敛、复杂度较低等特性。多径条件下随机测试仿真,以误差1度以内作为目标,能在两条相干径下达到87%的估计成功概率,三条相干径下达到83%的成功概率。本文选用AD9361+FMCCOMMS5的软件无线电平台模拟建立BLE测向环境,经过对多径环境下接收到测试数据分析发现本论文的算法能够在30度到150度范围内解相干,对来波估计平均误差在3度左右。因该算法复杂度()较空间平滑MUSIC算法复杂度(~2+~3+1)小一个数量级,所以对BLE定位解决多径问题具有一定的参考意义。更多还原

【Abstract】 In recent years,with the rapid development of the Internet of Things and artificial intelligence,positioning technology has also gained rapid advancement.At the same time,the demand for location technology is also increasing.However,the new indoor positioning techniques Which based on AoA(Angle of Arrival)or AoD(Angle of Depature)of Bluetooth Low Energe(BLE)have affected the positioning accuracy due to problems such as multipath interference.In response,a closed-form decoherence anti-multipath algorithm for BLE indoor positioning direction finding is implemented to solve the multipath interference problem affecting BLE positioning.In this thesis,based on the BLE signal transmission model and array direction finding theory,the BLE AoA/AoD direction finding model is being analysed firstly,and it is found that there is the multipath interference leading to large and unstable deviation of the direction finding angle.Therefore,this thesis simplifies the indoor positioning model to a model with mainly a direct path and a strong reflective interference path,and a decoherence algorithm for the closed-form solution of the fourantenna linear array is proposed.The main principle of the algorithm is to normalize the data received by the second,third and fourth antennas by conjugate multiplication with the first antenna,and simplify it into a three-dimensional linear equation group,which has a unique closed-form solution.the actual mathematical meaning of the algorithm is to solve for the first column of the incident signal covariance matrix.In addition,according to the principle that the signal amplitude of the general multipath reflection path is less than that of the direct path signal,the incident angle of the direct signal and the incident angle of the multipath reflection signal in the closed solution are distinguished by judging the amplitude of the coherent signal.Finally,after verifying the effectiveness of the implementation of the algorithm in the case of two multipath incidences,the thesis tests the ability of the algorithm to solve coherent vectoring under multiple multipath conditions and performs the algorithm complexity analysis to verify the algorithm’s advantage of stable vectoring and low complexity.The simulation results in this thesis show that the BLE AoA / AoD estimation of the four-antenna array by this algorithm has the characteristics of distinguishable multipath,certain convergence of solution and low complexity.Simulation results show that the random test error of simulation under multipath conditions is within 1 degree,and the estimated success probability can reach 87% under two coherent paths,and 83% under three coherent paths.In this thesis,the software defined radio platform of AD9361 + FMCCOMMS5 is used to simulate the BLE direction finding environment.Through the analysis of the received test data in multipath environment,it is found that the algorithm in this thesis can decoherence in the range of 30 degrees to 150 degrees,and the average error of the incoming wave estimation is about 3 degrees.The complexity of the algorithm O(ML)is one order of magnitude smaller than that of the Spatial Smoothing MUSIC algorithm O(M^2 L+M^3+1),so it has certain reference significance for BLE positioning to solve multipath problems.更多还原