CN106793087A - A kind of array antenna indoor positioning algorithms based on AOA and PDOA - Google Patents

Abstract

The present invention relates to a kind of array antenna indoor positioning algorithms based on AOA and PDOA, including：Multipath signal is distinguished using array antenna；Energy size according to the signal for receiving selects two most strong signal paths of energy, and the phase information of two signal paths is obtained from antenna；The angle of arrival of two bars is calculated using AOA methods of estimation；The propagation distance of two bars is calculated using PDOA methods of estimation；According to the position of the total barrier of positioning scene, virtual base station can be set up, non line of sight NLOS paths are converted into sighting distance LOS path；The position coordinates of object to be positioned is calculated using weighted least-squares WLS algorithms；Using residual weighted LS algorithms to obtaining more accurate position coordinates.

Description

A kind of array antenna indoor positioning algorithms based on AOA and PDOA

Technical field

The invention belongs to indoor positioning algorithms technical field, particular for multipath and the indoor positioning environment of non line of sight.

Background technology

The research and application of location technology are quite varied, permeate in various aspects, global location such as military, business, lives System (GPS) provides good Global Positioning Service, but it cannot be used under environment indoors.Passive ultra-high frequency radio frequency is known (UHF RFID) indoor positioning technologies are not paid close attention to by increasing people, because passive label has small volume, deployment letter It is single, without power supply, it is cheap, read distant, the advantage such as read or write speed, passive ultra-high frequency RFID location technology faster Context aware field is progressively applied to, such as in terms of logistics management, intelligent transportation, robot, office automatic.

The method of indoor wireless positioning includes being based on arrival time (TOA), reaching time-difference (TDOA), angle of arrival (AOA), arriving signal intensity (RSS) and arrival phase difference (PDOA) etc..However, in the UHF FRID systems of smaller bandwidth, The method of TOA and TDOA is difficult to be implemented；Signal fadeout of the method based on RSS to multipath is very sensitive, so based on simple The alignment system of RSS methods, positioning precision is not very high；Method based on PDOA is also influenceed by multipath so that range finding knot It is really inaccurate；Meanwhile, the method based on AOA is influenceed by non line of sight (NLOS) and multipath.

Common decrease multipath and non line of sight are divided into statistics method and geometry to the method that indoor position accuracy influences Method.No matter this two classes method is in the modeling link of front end, or location algorithm link in rear end suffers from being widely applied. For the method that modeling link suppresses error, statistical modeling is modeled generally be directed to non-market value, counts special with certain Levy to describe non-market value, or scattering object to causing multi-path problem is modeled；And Geometric Modeling, such as ray trace (Ray-tracing), mainly use radio signal propagation characteristic close with the propagation characteristic of light, signal Analysis are in specific environment The path of lower propagation.For the method that location algorithm link suppresses error, statistical method is missed by measuring reception signal Difference cloth is come to determine signal path be sighting distance or non line of sight；And geometric method is that signal road is determined by geometrical relationship Footpath is sighting distance or non line of sight.

The maximum shortcoming of statistical method is to need known signal statistical property under certain circumstances, and this is accomplished by passing through Substantial amounts of reality measures to obtain this statistical property.When the environment changes, statistical property can change, it is necessary to again a large amount of Measurement carry out corrected parameter.And geometric method does not need substantial amounts of measurement, known indoor arrangement figure is only needed.But it is existing Some methods are positioned using direct path or primary scattering path, are not deposited for Multiple Scattering problem, or hypothesis , or filtered by algorithm judgement.

The content of the invention

The present invention provides a kind of influence that can well improve multipath and non line of sight to indoor positioning, effectively improves and determine The indoor positioning algorithms of position precision, technical scheme is as follows：

A kind of array antenna indoor positioning algorithms based on AOA and PDOA, comprise the following steps：

1) multipath signal is distinguished using array antenna；

2) the energy size according to the signal for receiving selects two most strong signal paths of energy, and from antenna The middle phase information for obtaining two signal paths；

3) angle of arrival of two bars is calculated using AOA methods of estimation；

4) propagation distance of two bars is calculated using PDOA methods of estimation；

5) according to the position of the total barrier of positioning scene, virtual base station can be set up, non line of sight NLOS paths is converted into Sighting distance LOS path；

6) according to step 3) in the angle information, the step 4 that obtain) in the range information that obtains, joint step 5) in set up Virtual base station, the position coordinates (x ', y ') of object to be positioned is calculated using weighted least-squares WLS algorithms:

Z₁=[x ', y ', x '²+y′²]^T=(G^TWG)^-1G^TWH

WhereinDiag { } represents diagonal matrix,

Z=[x, y, x²+y²]^T, X_k=[x_k,y_k]^T, k=1,2 ..., N is corresponding with kth signal path virtual Base station, d_kIt is the path length of the kth signal path estimated by PDOA, and d_k≤d_k+1, θ_kIt is kth signal path Angle of arrival；

7) using residual weighted LS algorithms to step 6) position that obtains optimizes, and obtains more accurate position coordinates (x,y)。

The present invention distinguishes multipath signal using aerial array, and obstructed path is converted into sighting distance using virtual base station Path.The most strong signal path of two energy is chosen, using AOA and PDOA algorithm for estimating, combines weighted least-squares and residual error Weighting algorithm calculates the position of object to be positioned.The influence to indoor positioning of multipath and non line of sight can well be improved, had Improve positioning precision in effect ground.

Brief description of the drawings

Fig. 1 is overall flow block diagram of the invention.

Fig. 2 is positioning scene schematic diagram of the invention.

Obstructed path is converted into Fig. 3 the schematic diagram of los path using virtual base station.

Fig. 4 is the analogous diagram of the positioning precision using the algorithm in the present invention, and the environment of emulation is a room of 10m*10m Son, in the environment of white Gaussian noise, using algorithm of the invention, positioning precision is less than 0.5m with 80% probability.

Specific embodiment

Below with uniform linear array, algorithm of the present invention is described in detail.

1) multipath signal is distinguished using array antenna.

2) the energy size according to the signal for receiving selects two most strong signal paths of energy, and from antenna The middle phase information for obtaining two signal paths.

3) angle of arrival of two bars is calculated using AOA methods of estimation, according to step 2) in the phase information that obtains, The angle of arrival of signal can be just calculated using the MUSIC algorithms based on space smoothing classical in AOA methods of estimation.

4) propagation distance of two bars, the following institute of formula of PDOA methods range finding are calculated using PDOA methods of estimation Show：

Wherein, c is the aerial spread speed of light, and Δ φ represents that signal travels to object to be positioned and produces from array antenna Raw phase difference, π=3.14, Δ f is the difference on the frequency between two signals of different frequency, and d is the biography of the signal being calculated Broadcast distance.

5) according to the position of the total barrier of positioning scene, virtual base station can be set up, by the conversion of non line of sight (NLOS) path It is sighting distance (LOS) path.As shown in figure 3, X_bIt is the position of array antenna, X_tIt is the position of object to be positioned, X_vIt is according to obstacle The virtual base station that the position that thing l sets up is set up.From virtual base station as viewed from object to be positioned, we will obtain a sighting distance Path.

6) according to step 3) in the angle information, the step 4 that obtain) in the range information that obtains, joint step 5) in set up Virtual base station, the position of object to be positioned is calculated using weighted least-squares (WLS) algorithm:

It is provided with being estimated by entity base station apart from d and angle, θ for N signal paths, X_k=[x_k,y_k]^T, k=1, 2 ..., N is the virtual base station corresponding with kth signal path, d_kIt is the kth signal path estimated by PDOA Path length, and d_k≤d_k+1, θ_kIt is the angle of arrival of kth signal path.With X=[x, y]^TRepresent the position of object to be positioned Put, then：

x sin(θ_i)-y cos(θ_i)=x_i sin(θ_i)+y_i cos(θ_i)

Write two formula above as matrix form, GZ=H can be obtained, wherein：

Z=[x, y, x²+y²]^T。

The position of object to be positioned is calculated using weighted least square algorithm, the initial position co-ordinates of object to be positioned are obtained (x′,y′)：

Z₁=[x ', y ', x '²+y′²]^T=(G^TWG)^-1G^TWH,

WhereinDiag { } represents diagonal matrix,

7) setUsing residual weighted (LS) algorithm To step 6) position that obtains optimizes, and obtains the position X=[x, y] of more accurate object to be positioned^T, wherein：

X′_i=[x '_i,y′_i]^TI=1,2 ..., N is the object to be positioned derived from the i-th signal path Position.

Claims (1)

1. a kind of array antenna indoor positioning algorithms based on AOA and PDOA, comprise the following steps：

1) multipath signal is distinguished using array antenna.

2) the energy size according to the signal for receiving selects two most strong signal paths of energy, and is obtained from antenna Take the phase information of two signal paths；

3) angle of arrival of two bars is calculated using AOA methods of estimation；

4) propagation distance of two bars is calculated using PDOA methods of estimation；

5) according to the position of the total barrier of positioning scene, virtual base station can be set up, non line of sight NLOS paths is converted into sighting distance LOS path；

6) according to step 3) in the angle information, the step 4 that obtain) in the range information that obtains, joint step 5) in the void set up Intend base station, the position coordinates (x ', y ') of object to be positioned is calculated using weighted least-squares WLS algorithms:

Z₁=[x ', y ', x '²+y′²]^T=(G^TWG)^-1G^TWH

WhereinDiag { } represents diagonal matrix,

$G=\left[\begin{array}{ccc}-2x_1& -2y_1& 1\\ .& .& .\\ .& .& .\\ .& .& .\\ -2x_N& -2y_N& 1\\ \sin \left({\mathrm{\θ}}_1\right)& \cos \left({\mathrm{\θ}}_1\right)& 0\\ .& .& .\\ .& .& .\\ .& .& .\\ \sin \left({\mathrm{\θ}}_N\right)& \cos \left({\mathrm{\θ}}_N\right)& 0\end{array}\right],H=\left[\begin{array}{c}{d}_1^2-(x_1^2+y_1^2)\\ .\\ .\\ .\\ d_N^2-(x_N^2+y_N^2)\\ x_1\sin \left({\mathrm{\θ}}_1\right)-y_1\cos \left({\mathrm{\θ}}_1\right)\\ .\\ .\\ .\\ x_N\sin \left({\mathrm{\θ}}_N\right)-y_N\cos \left({\mathrm{\θ}}_N\right)\end{array}\right],$

Z=[x, y, x²+y²]^T, X_k=[x_k,y_k]^T, k=1,2 ..., N are the virtual base stations corresponding with kth signal path, d_kIt is the path length of the kth signal path estimated by PDOA, and d_k≤d_k+1, θ_kIt is arriving for kth signal path Up to angle；

7) using residual weighted LS algorithms to step 6) position that obtains optimizes, obtain more accurate position coordinates (x, y)。