JP7239958B2 - Self-adaptive indoor fusion positioning method based on dynamic environment - Google Patents

Description

TECHNICAL FIELD The present invention relates to the technical field of indoor positioning (positioning) and 5G, and in particular to a dynamic environment-based self-adaptive indoor fusion positioning method.

Along with the continuous development of communication technology and Internet technology, many technical fields have entered the location service industry together, and highly accurate location information is the foundation for providing high quality location services. At present, technical fields such as smart home, car navigation and mobile phone tracking have become hot research directions of wireless positioning technology. On the other hand, the Global Navigation Satellite System can provide users with relatively high-precision positioning services, and can basically meet users' needs based on location services in outdoor scenes. However, indoors (indoors), due to the shielding of buildings, the attenuation of GNSS signals is very fast and cannot be applied accurately for indoor positioning. Therefore, accurate positioning to the indoor environment has become a hotspot of positioning research, and the research results not only bring enormous economic benefits, but also can be further innovated by collaborating with other technical fields.

Current indoor positioning technologies include bluetooth technology, UWB technology, infrared technology, geomagnetic technology, zig-bee technology, etc., but all of these technologies have more or less shortcomings and are subject to changing indoor environments. It is difficult to maintain accuracy inside. On the other hand, indoor positioning based on the multilateration method is a common positioning method that determines signal attenuation according to the strength of the received signal and calculates the distance, but the distance determined by this method is There is a slight difference from the actual distance. In addition, the indoor positioning based on the CIR fingerprint positioning method is a database matching method, but the collected database cannot contain all the conditions of the indoor environment, so database matching errors will occur. It should be noted that the above techniques are further affected by changes in the room temperature, resulting in a decrease in positioning accuracy. With the development of communication technology and people's demand for wireless communication speed, the development of communication networks will accelerate, and cooperative positioning by 5G positioning base stations can be easily realized. In addition, people are constantly making innovative proposals for methods of fusion positioning. However, most indoor positioning do not analyze the impact of indoor environment changes on positioning, and lose self-adaptation to the environment. Also, the single room positioning method is always slightly less accurate than the fusion positioning method.

As mentioned above, most of the existing indoor positioning techniques cannot be corrected for indoor environmental changes, and the performance of single indoor positioning method is inferior to the fusion positioning method.

The present invention has been made in view of the problems in the above existing technology, and by providing a self-adaptive indoor fusion positioning method based on a dynamic environment, the positioning accuracy is improved and self-adaptive by image recognition. The purpose is to modify it and make it more adaptable to the environment.

To achieve the above objects, the dynamic environment-based self-adaptive indoor fusion positioning method according to the present invention comprises:
(1) obtaining the parameters of the positioning spot in the offline stage based on the CIR;
(2) obtaining and analyzing the parameters of the positioning target pending measurement based on the 5G positioning base station;
(3) positioning the positioning target pending measurement using a multilateration method and analyzing the error between the predicted position and the actual position;
a step (4) of locating the positioning target awaiting measurement using the CIR fingerprint positioning method and analyzing the error between the predicted position and the actual position;
The position measured using the multilateration method is defined as the measured value, the position predicted using the CIR fingerprint positioning method is defined as the predicted value, and the error between the measured value and the predicted value is a fusion calculation based on the mean square error. and estimating the best position (5);
a step (6) of self-adaptive fusion positioning according to changes in the indoor environment;
including.

Furthermore, the above step (2) is
(21) collecting parameters of time delay, power and frequency at a positioning target point pending measurement by a 5G positioning base station;
collecting and analyzing indoor environment images (22);
a step (23) of processing the acquired data;
including.

が得られ、ここで、ε（ｄ）≦ε_ｍａｘは距離誤差であり、第ｉ番目に計算する位置がｌ_ｄ＝（ｘ^ｉ _ｄ，ｙ^ｉ _ｄ）であり、ターゲットの実際の位置がｌ_ａ＝（ｘ^ｉ _ａ，ｙ^ｉ _ａ）であるとすると、平均誤差は以下の式

で表示でき、ここで、Ｊはトータルのデータを取得した回数を表す。 Furthermore, the above step (3) is realized by the method shown below,
Let d ₁ , d ₂ and d ₃ be the calculated distances between the three 5G positioning base stations and the target respectively, and (A ₁ , A ₂ ), (B ₁ , B ₂ ) and (C ₁ , C ₂ ) are Assuming that each represents the position of a 5G positioning base station, the following equation

where ε(d)≦ε _max is the range error, the i-th computed position is l _d =(x ⁱ _d , y ⁱ _d ) and the actual position of the target is l Assuming _a = (x ^ia _, y ^ia ₎ , the average error is given by

where J represents the number of times the total data was acquired.

で表され、ここで、ａ_ｎはチャネルゲイン、ｅ^{－ｊ２πｆτｎ}は周波数オフセット、τ_ｎはチャネル時間遅延を示し、
ＣＩＲ指紋データベースは以下の式

で表され、ここで、ｌ_ｃ＝（ｘ_ｃ，ｙ_ｃ）は位置を表し、Ｆはマッピング関係を示し、τ＝｛τ_１，τ_２，・・・τ_Ｎ｝は時間遅延を表し、ｐ＝｛ｐ_１，ｐ_２・・・ｐ_Ｎ｝は電力を表し、ｆ＝｛ｆ_１，ｆ_２・・・ｆ_Ｎ｝は周波数を表し、
仮に、第ｉ番目に予測された位置がｌ_ｃ＝（ｘ^ｉ _ｃ，ｙ^ｉ _ｃ）であり、ターゲットの実際の位置がｌ_ａ＝（ｘ^ｉ _ａ，ｙ^ｉ _ａ）である場合、平均誤差は以下の式

で表示できる。 Furthermore, the above step (4) is realized by the method shown below,
The 5G positioning base station builds an offline CIR fingerprint database based on the parameters of time delay, power and frequency, and the CIR is given by the following formula

where a _n is the channel gain, e ^−j2πfτn is the frequency offset, _τn is the channel time delay, and
The CIR fingerprint database has the following formula

where l _c =(x _c , y _c ) denotes the position, F denotes the mapping relation, τ={τ ₁ , τ ₂ , . . . τ _N } denotes the time delay, p={p ₁ , p _{2 . .} . p _N } represents power, f={f ₁ , _{f 2 .}
If the i-th predicted position is l _c =(x ⁱ _c ,y ⁱ _c ) and the actual position of the target is l _a =(x ⁱ _a ,y ⁱ _a ), then the average error is the following formula

can be displayed with

に示されるようにＡとＢとの両点間の受信信号の強度を得て、ここで、ＡおよびＢは二つの位置であり、ＱおよびＵは経験則に基づく定数であり、ＲＳＳＩ_ＡＢは位置Ａにおいて受信した位置ＢのＲＳＳＩ値を表し、環境変化はＱおよびＵの変化を引き起こし、二つの定数を推定することにより環境変化による影響を低減して環境変化の距離計算に対する影響が小さくなることを保証し、

であり、同様にＵ_Ｂ、Ｑ_ＢおよびＵ_Ｃ、Ｑ_Ｃを算出し、環境変化後のモデルパラメータは

に更新され、変化の最も小さい５Ｇ位置決め基地局を見つけ、画像認識と組み合わせて障害物の５Ｇ位置決め基地局に対する影響を判断し、障害物が５Ｇ位置決め基地局とターゲットとのパス間の内半平面内にある場合、ＣＩＲ信号に対する影響は大きく、パス外平面の５Ｇ位置決め基地局に切り替える必要があり、受信されたパラメータの相似性に基づいて、以下の式

により位置を予測し、ここで、［τ_ｅ，ｐ_ｅ，ｆ_ｅ］は環境変化後の受信パラメータを表し、［τ_ｋ，ｐ_ｋ，ｆ_ｋ］はオフラインＣＩＲ指紋データベースパラメータを表し、｜・｜は剰余演算を表し、
各５Ｇ位置決め基地局はそれぞれマルチラテレーション方法およびＣＩＲ指紋による位置決め方法の誤差修正データベースを構築し、修正後の誤差は、

であり、ここで、Ｃ_１およびＣ_２はそれぞれマルチラテレーション方法およびＣＩＲ指紋による位置決め方法の誤差修正のマッピング関係であり、続いて、以下の式

で平均二乗誤差を演算し、ここで、Ｈ_１およびＨ_２は誤差係数を表し、続いて、以下の式

で融合位置決めを演算し、ここで、ｌ_ａは融合位置決めの演算結果である。 Furthermore, the above step (6) is realized by the following method,
With the RSSI distance measurement model, the following equation

, where A and B are two locations, Q and U are empirical constants, and RSSI _AB is Represents the RSSI value of location B received at location A, environmental changes cause changes in Q and U, and two constants are estimated to reduce the impact of environmental changes to lessen the impact of environmental changes on range calculations. guarantee that

and similarly calculate U _B , Q _B and U _C , Q _C , and the model parameters after the environmental change are

to find the 5G positioning base station with the smallest change, combine with image recognition to determine the impact of obstacles on the 5G positioning base station, and determine if the obstacle is the inner half plane between the paths between the 5G positioning base station and the target If it is within, the impact on the CIR signal is large, it is necessary to switch to the 5G positioning base station in the out-of-path plane, based on the similarity of the received parameters, the following formula

where [τ _e , p _e , f _e ] represent the reception parameters after the environmental change, [τ _k , p _k , f _k ] represent the offline CIR fingerprint database parameters, | | represents the remainder operation,
Each 5G positioning base station builds an error correction database of the multilateration method and the positioning method by CIR fingerprint, and the error after correction is

where C ₁ and C ₂ are the error correction mapping relations of the multilateration method and the CIR fingerprint localization method, respectively, followed by the following equation

where H ₁ and H ₂ represent the error coefficients, followed by the following equation

where _la is the result of the fused positioning.

Compared with the existing technology, the present invention has the following beneficial effects.
According to the present invention, by measuring indoor changes, 5G positioning base station switching and error correction can be performed, so that the positioning model can adapt to environmental changes and improve prediction accuracy.

1 is a flowchart of a method according to the invention; It is a figure for demonstrating the multilateration method. FIG. 4 is a diagram for explaining a positioning method using CIR fingerprints; FIG. 4 is a diagram for explaining switching of base stations by the multilateration method; FIG. 4 is a diagram for explaining switching of base stations by a positioning method based on CIR fingerprints;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments according to the present invention will be described in detail with reference to the drawings.
As shown in FIG. 1, according to the dynamic environment-based self-adaptive indoor fusion positioning method of the present invention, firstly, the characteristics of each spot in the room are acquired, and the positioning is performed by the 5G positioning base station; , error analysis on the difference from the actual position when using the CIR fingerprint localization method and the multilateration method. Furthermore, changes in the surrounding environment are identified by image recognition, and base station switching and error correction are performed by feeding back the identification results to the fusion system. In this way, more accurate and reliable position prediction with high environmental adaptability can be performed.

The present invention mainly includes the following three contents. The first content is the error analysis of the positioning methods, analyzing the average error of the two methods by comparing the calculated position with the actual position. The second content is to use image recognition to identify environmental changes and exclude 5G positioning base stations that are seriously affected for positioning. Then, the identification result is used to update the offline CIR fingerprint database of the 5G positioning base station that is seriously affected, correct the error, and realize the overall update. The third content is to use the fusion localization method to predict the optimal position of the target. Specifically, it includes the following steps.

(Step 1)
Obtain the parameters of the positioning spot in the off-line stage according to the CIR. and obtain and analyze the parameters of the positioning target waiting for measurement according to the 5G positioning base station.

First, the 5G positioning base station collects parameters such as time delay, power and frequency in the positioning target spot pending measurement. Images of the indoor environment are then collected and analyzed. Subsequently, the acquired data is processed.

(Step 2)
A multilateration method is used to locate the measurement-pending positioning targets, and the errors between the predicted and actual positions are analyzed.

が得られる。ここで、ε（ｄ）≦ε_ｍａｘは距離誤差であり、上式に解があることを保証する。 (1) Error analysis by multilateration method As shown in Figure 2, because the channel changes over time, the distance measured by each 5G positioning base station has a certain range, and the three 5G positioning with the highest signal strength When the position is calculated by the base station, the area where the transform ranges overlap is the position of the target, and a certain error may occur. Let d ₁ , d ₂ and d ₃ be the calculated distances between the three 5G positioning base stations and the target respectively, and let (A ₁ , A ₂ ), (B ₁ , B ₂ ) and (C ₁ , C _{2 )} ) represents the location of the 5G positioning base station, then the following equation

is obtained. where ε(d) ≤ ε _max is the distance error, ensuring that the above equation has a solution.

で表示でき、Ｊはトータルのデータを取得した回数を表す。 If the i-th computed position is l _d =(x ⁱ _d ,y ⁱ _d ) and the actual position of the target is l _a =(x ⁱ _a ,y ⁱ _a ), then the average error is the following formula

and J represents the number of times the total data was obtained.

で表され、ここで、ａ_ｎはチャネルゲイン、ｅ^{－ｊ２πｆτｎ}は周波数オフセット、τ_ｎはチャネル時間遅延を示す。 (Step 3)
The CIR fingerprint positioning method is used to position the positioning target pending measurement, and the error between the predicted position and the actual position is analyzed.
First, many 5G positioning base stations build offline CIR fingerprint databases (each 5G positioning base station's offline CIR fingerprint database is basically different) based on parameters such as time delay, power and frequency. As shown in FIG. 3, shaded areas represent spots that are very close in distance, the CIR signal in that environment is essentially the same, and the channel varies over time. Although this method may have a certain estimation error, CIR takes into account the multipath effect, so the above parameters can give a unique position, and the accuracy is relatively high. CIR is the following formula

where a _n is the channel gain, e ^−j2πfτn is the frequency offset, and τ _n is the channel time delay.

で表され、ここで、ｌ_ｃ＝（ｘ_ｃ，ｙ_ｃ）は位置を表し、Ｆはマッピング関係を示し、τ＝｛τ_１，τ_２，・・・τ_Ｎ｝は時間遅延を表し、ｐ＝｛ｐ_１，ｐ_２・・・ｐ_Ｎ｝は電力を表し、ｆ＝｛ｆ_１，ｆ_２・・・ｆ_Ｎ｝は周波数を表す。 The CIR fingerprint database has the following formula

where l _c =(x _c , y _c ) denotes the position, F denotes the mapping relation, τ={τ ₁ , τ ₂ , . . . τ _N } denotes the time delay, p={p ₁ , p _{2 . .} . p _N } represents power and f={f ₁ , _{f 2} .

で表される。 If the i-th predicted position is l _c =(x ⁱ _c ,y ⁱ _c ) and the actual position of the target is l _a =(x ⁱ _a ,y ⁱ _a ), then the average error is the following formula

is represented by

(Step 4)
Fusion positioning is achieved based on the mean squared error.
First, by combining the multilateration method and the CIR fingerprint positioning method, each time position prediction is performed, the position measured using the multilateration method is used as the measured value, and the CIR fingerprint positioning method is used for prediction. Then, the error between the measured value and the predicted value is fusion calculated based on the mean square error, and finally the optimum position is estimated.

(Step 5)
Fusion positioning is performed in a self-adaptive manner according to changes in the indoor environment.
Before positioning, the indoor environment is identified by image recognition to ensure that the system adjusts itself adaptively. Changes in the indoor environment can be quickly reflected by comparing the previous image (one time before) and the current image (current time) and combining it with a neural network algorithm. Some large changes, such as the movement of indoor furniture and the movement of people, cause large errors in both of the two types of positioning algorithms described above. need to raise.

に示されるようにＡとＢとの両点間の受信信号の強度を得ることができる。ここで、ＡおよびＢは二つの位置であり、ＱおよびＵは経験則に基づく定数であり、ＲＳＳＩ_ＡＢは位置Ａにおいて受信した位置ＢのＲＳＳＩ値を表す。環境変化はＱおよびＵの変化を引き起こすため、二つの定数を推定することにより環境変化による影響を低減して環境変化の距離計算に対する影響が小さくなることを保証し、

である。 For the multilateration method, the three 5G positioning base stations with the strongest signal strength are used to calculate the position of the target. Therefore, when the environment changes, it spontaneously switches the 5G positioning base station, as shown in FIG. With the RSSI distance measurement model, the following equation

The strength of the received signal between points A and B can be obtained as shown in . where A and B are two locations, Q and U are heuristic constants, and RSSI _AB represents the RSSI value of location B received at location A. Since environmental changes cause changes in Q and U, estimating two constants reduces the impact of environmental changes to ensure that environmental changes have less impact on the distance calculation;

is.

に更新される。 Similarly, _UB , _QB and _UC , _QC are calculated. Then, the model parameters after the environmental change are

is updated to

により位置を予測する。ここで、［τ_ｅ，ｐ_ｅ，ｆ_ｅ］は環境変化後の受信パラメータを表し、［τ_ｋ，ｐ_ｋ，ｆ_ｋ］はオフラインＣＩＲ指紋データベースパラメータを表し、｜・｜は剰余演算を表す。 For the CIR fingerprint positioning method, the off-line CIR fingerprint database obtained using a 5G positioning base station with no obstacles within line-of-sight is used. According to the above explanation, each indoor 5G positioning base station builds its own offline CIR fingerprint database in the offline stage, but when the environment changes due to obstacles, the multipath effect experienced by each 5G positioning base station also changes. Therefore, it is necessary to find the 5G positioning base station with the smallest change, and combine with image recognition to determine the impact of obstacles on the 5G positioning base station. As shown in FIG. 5, if the obstacle is in the inner half-plane between the path between the 5G positioning base station and the target, the impact on the CIR signal is large, and it is necessary to switch to the 5G positioning base station in the out-of-path plane. This is because the out-of-path plane has less influence on the 5G positioning base station. Then, based on the similarity of the received parameters, the following formula

Predict the position by where [τ _e , p _e , f _e ] represent the reception parameters after the environmental change, [τ _k , p _k , f _k ] represent the offline CIR fingerprint database parameters, |·| .

For the 5G positioning base stations that meet the out-of-path plane, the influence of the external environment is large, so it is necessary to dynamically update the offline CIR fingerprint database of these instruments, specifically including the following steps.

As the first step, image recognition is used to determine changes in the indoor environment, and based on the 5G positioning base stations that meet the conditions of the plane inside and outside the path, the 5G positioning base stations are distributed, and the 5G positioning conditions within the path are determined. For the base station, it is determined that the offline CIR fingerprint database needs to be updated, and the 5G positioning base station with the strongest received signal is found from the set of 5G positioning base stations that satisfy the out-of-path conditions to perform prediction.

ここで、Ｆ｛｝は更新前のマッピング関係を示し、Ｆ’｛｝は更新後のマッピング関係を示し、ｌ_ｐはｌ_ｃに対する修正位置であって、パス外の条件を満たす５Ｇ位置決め基地局に基づいて予測した結果である。 As a second step, update the offline CIR fingerprint database based on the 5G positioning base station and prediction results, as shown in the following formula.

Here, F{ } indicates the mapping relationship before updating, F'{ } indicates the mapping relationship after updating, l _p is the corrected position for l _c , and the 5G positioning base station that satisfies the out-of-path condition. This is the result predicted based on

であり、ここで、Ｃ_１およびＣ_２はそれぞれマルチラテレーション方法およびＣＩＲ指紋による位置決め方法の誤差修正のマッピング関係である。 Then, error correction is performed. Because image recognition identifies environmental changes, 5G positioning base station switching can improve positioning accuracy to some extent, but environmental changes also affect the previously analyzed positioning error. Therefore, it is necessary to correct errors in the positioning method in consideration of identification by image recognition. Of course, the positioning error influencing factors include the new obstacle type, size and distance between the new obstacle and the 5G positioning base station. Therefore, image recognition analyzes the type and size of new obstacles, and 5G base station positioning technology analyzes the distance between new obstacles and itself. Each 5G positioning base station builds an error correction database of the multilateration method and the positioning method by CIR fingerprint, and the error after correction is

where C ₁ and C ₂ are the error correction mapping relations of the multilateration method and the CIR fingerprint localization method, respectively.

When it is identified that the environment has changed by image recognition, after switching the base station, error correction is performed not only for the base station currently positioned, but also for the base station for positioning.

Two kinds of positioning methods are implemented and switching is self-adaptive to environmental changes, but there is a certain error in the single positioning method, so the fusion positioning method is adopted.

で平均二乗誤差を演算し、ここで、Ｈ_１およびＨ_２は誤差係数を表す。続いて、以下の式

で融合位置決めを演算し、ここで、ｌ_ａは融合位置決めの演算結果である。 As known, the position calculated by the multilateration method is l _d =(x _d , y _d ), the correction error is e _nda , and the position predicted by the CIR fingerprint localization method is l _c =(x _c , y _c )=F′{τ,p,f} and the correction error is en _da , so the method of mean squared error is used to perform fusion positioning. For the sake of generality, if _xc =max{ _xd , _yd } and _yd =max{ _yc , _yd }, the calculation flow is as follows. the formula below

where H ₁ and H ₂ represent the error coefficients. Then the following formula

where _la is the result of the fused positioning.

Claims (4)

(2) obtaining, analyzing and processing the parameters of the positioning target pending measurement based on the 5G positioning base station;
the step (3) of locating the positioning target awaiting measurement using a multilateration method and analyzing the error between the predicted position and the actual position;
step (4) of locating the positioning target awaiting measurement using the CIR fingerprint parameter database built from the parameters processed in step (2) and analyzing the error between the predicted and actual positions;
The position measured using the multilateration method is defined as the measured value, the position predicted using the CIR fingerprint positioning method is defined as the predicted value, and the error between the measured value and the predicted value is a fusion calculation based on the mean square error. and estimating the best position (5);
a step (6) of self-adaptive fusion positioning according to changes in the indoor environment;
including
The step (6) is realized by the following method,
With the RSSI distance measurement model, the following equation

, where A and B are two locations, Q and U are empirical constants, and RSSI _AB is Represents the RSSI value of location B received at location A, environmental changes cause changes in Q and U, and two constants are estimated to reduce the impact of environmental changes to lessen the impact of environmental changes on range calculations. guarantee that

and similarly calculate U _B , Q _B and U _C , Q _C , and the model parameters after the environmental change are

, find the 5G positioning base station with the smallest change, combine with image recognition to determine the impact of obstacles on the 5G positioning base station, and determine if the obstacle is within the line- of-sight distance between the path between the 5G positioning base station and the target If it is in the half-plane, the impact on the CIR signal is large, it is necessary to switch to the 5G positioning base station in the line-of-sight path out-of-plane, based on the similarity of the received parameters, the following formula

where [τ _e , p _e , f _e ] represent the reception parameters after the environmental change, [τ _k , p _k , f _k ] represent the offline CIR fingerprint database parameters, | | represents the remainder operation,
Each 5G positioning base station builds an error correction database of the multilateration method and the positioning method by CIR fingerprint, and the error after correction is

where C ₁ and C ₂ are the error correction mapping relations of the multilateration method and the CIR fingerprint localization method respectively, and 'category' is the 'new obstacle type' and ' size” is the “size of the new obstacle”, “distance ₁ ” and “distance ₂ ” are the “distances between the new obstacle and the 5G positioning base station”, and “e _da ” and “e _ca ” is the “average error” followed by the following formula

where H ₁ and H ₂ represent the error coefficients, followed by the following equation

where l _a is the result of the calculation of the fusion positioning, x _a and y _a represent the abscissa and ordinate obtained by calculating the fusion positioning, and x _d and y _d denote the abscissa and ordinate of the position calculated by the multilateration method, and x _c and y _c denote the abscissa and ordinate of the predicted position by the CIR fingerprint localization method. A self-adaptive indoor fusion positioning method based on a dynamic environment. The step (2) is
collecting parameters of time delay, power and frequency at the positioning target spot awaiting measurement by the 5G positioning base station (21);
collecting and analyzing images of the indoor environment (22);
a step (23) of processing the acquired data;
The dynamic environment-based self-adaptive indoor fusion positioning method according to claim 1, characterized by comprising: The step (3) is realized by the method shown below,
Let d ₁ , d ₂ and d ₃ be the calculated distances between the three 5G positioning base stations and the target respectively, and (A ₁ , A ₂ ), (B ₁ , B ₂ ) and (C ₁ , C ₂ ) are Assuming that each represents the position of a 5G positioning base station, the following equation

where ε(d)≦ε _max is the range error, the i-th computed position is l _d =(x ⁱ _d , y ⁱ _d ) and the actual position of the target is l Assuming _a = (x ^ia , y ^ia ), _the average error e _{da is} given by

The dynamic environment based self-adaptive indoor fusion positioning method according to claim 1, characterized in that it can be represented by J, where J represents the number of times the total data has been acquired. The step (4) is realized by the method shown below,
In the CIR fingerprint parameter database constructed by the parameters processed in step (2) , the CIR is given by the following formula:

where a _n is the channel gain, e ^−j2πfτn is the frequency offset, _τn is the channel time delay, and
The CIR fingerprint database has the following formula

where l _c =(x _c , y _c ) denotes the position, F denotes the mapping relation, τ={τ ₁ , τ ₂ , . . . τ _N } denotes the time delay, p={p ₁ , p _{2 . .} . p _N } represents power, f={f ₁ , _{f 2 .}
If the i-th predicted position is l _c =(x ⁱ _c ,y ⁱ _c ) and the actual position of the target is l _a =(x ⁱ _a ,y ⁱ _a ), then the average error is the following formula

The dynamic environment-based self-adaptive indoor fusion positioning method according to claim 1, characterized by: .

Images