CN-121980313-A - RFID pose estimation non-line-of-sight shielding influence suppression method based on antenna weights

Abstract

The invention relates to the technical field of wireless sensing and discloses an RFID pose estimation non-line-of-sight shielding influence suppression method based on antenna weights, which comprises the following steps of S1, constructing a relation model of tag positions and non-line-of-sight shielding errors, S2, distributing antenna weights according to theoretical phase differences and actual phase differences of antennas and reference antennas, S3, constructing an augmentation Lagrange convex optimization algorithm according to the relation model of S1 and the antenna weights of S2, suppressing the influence of the non-line-of-sight shielding through continuous iteration, and finally obtaining tag position information, and S4, combining the antenna weights and a multi-signal classification algorithm, and obtaining tag pose information through peak detection. In addition, the method automatically identifies the blocked situation of the antenna and reduces the corresponding weight through an adaptive antenna weight distribution mechanism based on phase residual error statistical analysis.

Inventors

• Meng Zhaozong
• ZHAO JIAYU
• GAO NAN
• Ni Yubo
• ZHANG ZONGHUA

Assignees

• 河北工业大学

Dates

Publication Date

20260505

Application Date

20260123

Claims (5)

1. 1. The RFID pose estimation non-line-of-sight shielding influence suppression method based on the antenna weight is characterized by comprising the following steps of: s1, constructing a relation model of the label position and the non-line-of-sight shielding error; s2, distributing antenna weights according to the theoretical phase difference and the actual phase difference between each antenna and the reference antenna; S3, constructing an augmented Lagrangian convex optimization algorithm according to the relation model of S1 and the antenna weight of S2, and finally obtaining label position information by continuously iterating to inhibit the influence of non-line-of-sight shielding; S4, combining antenna weights and a multi-signal classification algorithm, constructing a weighted covariance matrix to inhibit non-line-of-sight shielding influence, and finally obtaining tag attitude information through peak detection.
2. 2. The method for suppressing influence of non-line-of-sight occlusion in RFID pose estimation based on antenna weights according to claim 1, wherein S1 comprises the steps of: The tag phase information is acquired by the reader, and the phase value phi can be expressed as: Wherein d represents the distance between the antenna and the tag, phi h 、φ p 、φ t represents the phase shift caused by the reader circuit, the tag gesture and the tag hardware, f represents the transmitting frequency of the antenna, and c represents the light speed; the method comprises the steps of placing a first antenna in a sight distance environment to serve as a reference antenna, expanding phi to obtain a measured distance D i from an ith antenna to a tag, and expressing the difference of square values of the distances between the ith antenna and the reference antenna as follows: Where (x, y) denotes the tag coordinates, (x i ,y i ) denotes the i-th antenna coordinates, Φ i,nlos denotes the phase error of the i-th antenna caused by non-line-of-sight occlusion, d i,nlos denotes the distance error of the i-th antenna caused by non-line-of-sight occlusion, d i denotes the true distance of the i-th antenna to the tag, L i =x2i+y2i; Defining D i,1 =D i -D 1 , wherein the relation between the label position and the non-line-of-sight occlusion error is as follows: the system has N antennae, and the relation model of the label position tau and the non-line-of-sight shielding error is as follows: In the formula, 。
3. 3. The method for suppressing influence of non-line-of-sight occlusion in RFID pose estimation based on antenna weights according to claim 1, wherein said S2 comprises the steps of: The method comprises the steps of obtaining initial rough estimated coordinates of a tag according to received phase values of all antennas by using a least square method, obtaining theoretical phase differences delta theta i,1 of an ith antenna and a reference antenna according to the initial rough estimated coordinates of the tag, calculating differences between the theoretical phase differences and actual phase differences, obtaining phase difference offset values delta phi i,1 , and distributing antenna weights omega according to distribution conditions P (delta phi i,1 |tau) of the phase difference offset values, wherein the antenna positions are fixed and known: Where the numerator of P (ΔΦ i,1 |τ) is a Gaussian probability formula, the denominator is a normalized coefficient, and μ i,1 and σ i,1 are the mean and standard values of the i-th antenna phase measurement residuals.
4. 4. The method for suppressing influence of non-line-of-sight occlusion in RFID pose estimation based on antenna weights according to claim 1, wherein said S3 comprises the steps of: Combining the relationship model and the antenna weights can obtain a weighted relationship model alpha: Taking the logarithm of the gaussian probability density function P (α|τ) of α gives the objective function F τ : in the formula, u i is the mean value of the measured phase of the ith antenna, the function is a convex function, when the minimum value is taken, the matching probability of the actual error distribution and the theoretical error distribution is maximum, at the moment, tau is the ideal coordinate of the tag, According to the convex function, an augmented Lagrangian convex optimization algorithm is constructed: Where λ is the Lagrangian multiplier, s is the relaxation variable, μ is the penalty parameter, h τ is the constraint vector, and X max 、Y max is the maximum abscissa and ordinate of the measurement scene, respectively.
5. 5. The method for suppressing influence of non-line-of-sight occlusion in RFID pose estimation based on antenna weights according to claim 1, wherein the S4 process is as follows: And carrying out differential operation on phase signals obtained by the same acquisition of the adjacent tag on the ith antenna to obtain a phase signal matrix x i (k) without position parameters: obtaining a signal covariance matrix R i according to the signal matrix: (10) Constructing a weighted covariance matrix R ω by combining antenna weights, and obtaining a characteristic vector R thereof; Wherein U N is a noise subspace formed by smaller eigenvalues of R ω , and U s is a signal subspace formed by larger eigenvalues of R ω ; obtaining a multi-signal classification spectrogram P MUSIC according to the weighted covariance matrix; ; a ω (θ) is a direction vector orthogonal to U N , and a peak point of P MUSIC is a tag attitude angle.

Description

RFID pose estimation non-line-of-sight shielding influence suppression method based on antenna weights Technical Field The invention relates to the technical field of wireless sensing, in particular to an RFID pose estimation non-line-of-sight shielding influence inhibition method based on antenna weights. Background Accurate acquisition of object grade and pose information in an industrial scene is a key support for informatization and intellectualization of an industrial process. The industrial Internet of things technology can integrate information such as object environment and identity by taking object pose information as a carrier, and can correlate the object pose information and the identity according to the position information of the object and the production unit, so that the perception range of the production module is enlarged, and the production efficiency of a production line is effectively improved by integrating the information and enlarging the perception range. Therefore, the indoor positioning technology based on the industrial scene becomes a novel technology which is paid attention to in the fields of industrial Internet of things and intelligent manufacturing. Part of traditional indoor positioning technology is limited by different factors, and accurate estimation of object grade and pose in complex industrial scenes is difficult to realize. For example, the visual positioning technology has large calculation amount and high equipment requirement, and seriously depends on illumination conditions, so that accurate tracking and positioning of objects are difficult to realize in industrial scenes with more non-line-of-sight shielding and complex light distribution. The inertial navigation positioning technology can generate irremovable offset and accumulated errors in the positioning process, and is difficult to meet the long-time positioning requirement in industrial scenes. Compared with the two methods, the radio wave positioning technology is more suitable for indoor positioning based on industrial scenes due to the characteristics of low computational complexity, strong environmental adaptability and non-contact multi-target positioning. Although the indoor positioning technology based on RFID has been significantly advanced under the push of emerging technologies such as intelligent manufacturing and the Internet of things; Indoor industrial scenes are usually limited in space and compact in structure, people flow frequently, the machine operation condition is complex, and non-line-of-sight shielding is easy to generate. Under such shielding, the radio frequency signals can generate serious attenuation, so that signal parameters acquired by a receiving end are distorted, and the pose calculation based on the distorted signals introduces larger systematic deviation, thereby remarkably reducing the pose estimation precision. In a dynamic positioning scene, non-line-of-sight errors are continuously accumulated and amplified in the process of continuously estimating the pose of a moving object, the precision of a positioning result is continuously reduced, and the continuous and stable pose estimation requirement in an industrial scene cannot be met. In addition, due to inherent characteristics of the RFID technology, the RFID signal contains parameter information such as the gesture, the position, the environment and the like of the target, and the parameters have strong coupling relation in a signal model. The traditional method is difficult to realize multi-parameter joint decoupling under a single observation frame, independent estimation is often needed, so that position and attitude information acquisition is asynchronous and mutually restricted, the overall performance and instantaneity of pose estimation are affected, and in view of the fact, the RFID pose estimation non-line-of-sight shielding influence inhibition method based on antenna weights is provided. Disclosure of Invention The invention aims to provide an RFID pose estimation non-line-of-sight shielding influence suppression method based on antenna weights, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the RFID pose estimation non-line-of-sight shielding influence inhibition method based on antenna weights comprises the following steps: s1, constructing a relation model of the label position and the non-line-of-sight shielding error; s2, distributing antenna weights according to the theoretical phase difference and the actual phase difference between each antenna and the reference antenna; S3, constructing an augmented Lagrangian convex optimization algorithm according to the relation model of S1 and the antenna weight of S2, and finally obtaining label position information by continuously iterating to inhibit the influence of non-line-of-sight shielding; S4, combining antenna weights and a multi-signal classificat