CN109738863A - A WiFi Fingerprint Indoor Localization Algorithm and System Based on Confidence - Google Patents

Abstract

The present invention proposes a WiFi fingerprint indoor positioning algorithm and system with integrated confidence, which includes the following steps: in the offline collection stage, the collection equipment is used to collect multiple AP signal strength characteristic values used for positioning at each reference point, and the signal strength The eigenvalues are preprocessed to calculate the mean value and mean square error of signal strength characteristics; after obtaining the mean square error, use the constructed virtual mapping table of mean square error interval and confidence degree to determine the confidence degree. , and the geographic location coordinates are uploaded to the server location fingerprint database together; in the online positioning stage, the real-time acquisition of the signal strength characteristic value data of the user to be located point and the characteristic data in the database are added with confidence in the Euclidean distance matching calculation, which can optimize the effect of distance. The invention uses the WiFi fingerprint indoor positioning optimization algorithm of fusion confidence, which can reduce jumping, increase stability and improve positioning accuracy.

Description

A WiFi Fingerprint Indoor Localization Algorithm and System Based on Confidence

technical field

The invention relates to the technical field of indoor positioning, in particular to a WiFi fingerprint indoor positioning algorithm and system integrating confidence.

Background technique

People spend more than 80% of their time in an indoor environment, but due to various factors such as technology and cost, indoor positioning does not have a wide range of applications. At present, scholars at home and abroad have proposed indoor positioning technologies and application systems based on wireless fidelity, Bluetooth, infrared, radio frequency identification, ultrasound, and ultra-wideband, taking into account factors such as deployment cost and positioning accuracy. Among them, the application of WiFi-based WLAN indoor positioning technology The prospects are the broadest.

RSS-based WLAN indoor positioning technology utilizes the correlation between RSS values and geographic location coordinates to achieve target location estimation. This technology has the following advantages: First, the deployment cost is low and the deployment is simple and quick. This technology uses the existing WLAN system and the portable mobile terminal in the user's hand to estimate the user's position through a positioning algorithm. At the same time, due to the high cost of the offline acquisition stage and in order to ensure the accuracy of the acquisition, a high-quality offline database will play a great role in the positioning accuracy. The inaccurate signal acquisition caused by the complex indoor environment, multipath effect, radio frequency interference and other factors will eventually lead to large online positioning errors.

Therefore, it is extremely important to design a relatively stable offline acquisition step, and the method of fusing confidence in the offline acquisition stage can further improve the positioning accuracy.

SUMMARY OF THE INVENTION

The purpose of the present invention is to propose a WiFi fingerprint indoor positioning algorithm and system integrating confidence, which can reduce jumping, increase stability, optimize Euclidean distance, and improve positioning accuracy.

In order to achieve the above object, one aspect of the embodiments of the present invention provides a WiFi fingerprint indoor positioning algorithm fused with confidence, including the following steps:

In the offline collection stage, the collection device is used to collect multiple AP signal strength characteristic values used for positioning at each reference point, and the signal strength characteristic values are preprocessed to calculate the signal strength characteristic mean value and mean square error;

After obtaining the mean square error, use the constructed virtual mapping table of mean square error interval and confidence level to determine the confidence level, and finally upload it to the server location fingerprint database together with the serial number of the reference point, the signal strength feature mean value set, and the geographic location coordinates;

In the online positioning stage, the confidence is added when the Euclidean distance matching calculation is performed between the signal strength characteristic value data of the point to be located in real time and each group of characteristic data in the database, that is, before the square sum is added in the calculation of the Euclidean distance. The reciprocal of the corresponding AP confidence, the geographic location coordinate corresponding to the reference point with the smallest Euclidean distance is used as the position estimate of the user's to-be-located point, and returned to the user, that is, the positioning is completed.

Further, the step of calculating the mean value of the signal strength feature includes:

According to the formula A single AP at a certain reference point collects n times of signal strength characteristic values and performs averaging processing to obtain an AP signal strength average value;

in, is the average signal strength of the AP, The sum of the signal strength eigenvalues collected n times for the AP.

According to the formula The signal strength characteristic values of all APs collected at each reference point are respectively averaged to obtain the average value of the signal strength of each AP at each reference point;

in, is the mean value of the signal strength feature of the mth AP at a certain reference point, is the sum of the signal strength characteristic values of the mth AP of a certain reference point collected n times, n is the collection times of the collection device, i is an independent variable, m is the number of APs, A set of average signal strengths for all APs collected at a single reference point, respectively, A set of sums of signal strength eigenvalues is obtained for all APs collected at a single reference point.

Further, the mean square error of the calculated signal strength eigenvalues is:

After each reference point receives the characteristic value of each AP signal strength, the mean value is processed and the mean square error is calculated at the same time;

The mean square error formula is

in, is the characteristic value of the signal strength collected by a single AP for the nth time, is the average signal strength of the AP.

Further, the calculation of the mean square error of the signal strength eigenvalues includes the following steps:

Each time the acquisition of the signal strength characteristic value of a reference point is completed, all APs calculate the mean square error separately and compare it with the threshold value.

If the mean square error is greater than the threshold, the calculated mean value and the collected signal strength eigenvalues are subtracted one by one to obtain the absolute value, and then sorted in descending order, the signal strength eigenvalues corresponding to the first three largest absolute values are eliminated, and the remaining signal strength eigenvalues are recalculated. The mean squared error is recorded, and it is also judged whether it is within the threshold until it meets.

Further, the judging whether the mean square error is within the threshold value further includes:

If the signal of an AP is not very stable, the mean square error is always greater than the threshold, then re-sampling;

If it is still greater than the threshold after re-sampling, the calculated mean value and the collected signal strength characteristic value are subtracted one by one to obtain the absolute value, and then sorted in descending order, and the signal strength characteristic value corresponding to the first three largest absolute values is eliminated. The mean square error of the intensity eigenvalues is recalculated and recorded until the last three signal intensity characteristic values are left, and the smallest mean square error and mean value in the whole process are recorded.

Further, the mean square error obtained by each AP of each reference point is converted to the mean square error according to the mean square error and confidence level mapping table to obtain the AP confidence level of the reference point; the mean square error and confidence level mapping table is a virtual mapping table, which can be Created in the program, it is divided into two types of elements: mean square error interval and confidence level.

Further, each AP confidence level obtained for each reference point and the geographic location coordinates, serial number, and signal strength feature mean value set of the reference point are uploaded to the server location fingerprint database together, wherein the signal strength feature mean value set in the database is recorded as : , N is the number of reference points, and m is the number of APs.

Further, in the online positioning stage, signal strength collection is first performed at a certain point to be positioned for a period of time. During this period of time, the signal strength characteristic values of all APs used for positioning will be collected, and the signal strengths of all APs will be collected. The eigenvalues are averaged respectively, and a set of AP signal strength feature averages will be obtained, which is recorded as: , m is the number of APs, upload the mean set to the server, and add confidence in the Euclidean distance matching calculation with each set of feature data in the database, that is, multiply the relative value before the square and add in the calculation of the Euclidean distance. The reciprocal of the corresponding AP confidence, the geographic location coordinate corresponding to the reference point with the smallest Euclidean distance is used as the position estimate of the point to be located by the user, and returned to the user, that is, the positioning is completed.

The original Euclidean distance formula is:

Let the confidence of the kth AP under the jth reference point in the database be , the Euclidean distance formula is changed to:

Where m is the number of APs, N is the total number of reference points in the database, A _k is the signal strength characteristic value of the k-th AP measured by the user at the point to be located, and B _jk is the k-th reference point under the j-th reference point in the database The signal strength characteristic value of each AP.

In order to achieve the above object, another aspect of the embodiments of the present invention also provides a WiFi fingerprint indoor positioning system with integrated confidence, including a mobile terminal, a cloud server that communicates with the mobile terminal network, and the mobile terminal includes a WiFi acquisition unit and a processing unit. In the program, the cloud server includes a WiFi characteristic parameter storage unit and a positioning unit.

The WiFi acquisition unit and processing program are used to collect multiple AP signal strength characteristic values received by each reference point in the offline collection stage, and perform mean value and mean square error processing on the signal strength characteristic values and record, wherein each reference Points include multiple AP signal strength eigenvalues.

The WiFi characteristic parameter storage unit is used to store the MAC address, confidence level, geographic location coordinates of each AP, and the characteristic mean value and mean square error of each AP's signal strength at each reference point.

The positioning unit is used in the online positioning stage, and the confidence is added when the Euclidean distance matching calculation is performed between the signal strength characteristic value data of the user to be positioned point acquired in real time and each group of signal strength characteristic value data in the database, that is, when calculating the Euclidean distance. Multiply the reciprocal of the confidence of the corresponding AP before the square and sum in the distance, so as to achieve the effect of optimizing the distance.

Compared with the prior art, the beneficial effects of the present invention are:

The invention uses the WiFi fingerprint indoor positioning optimization algorithm fused with confidence, which can reduce jumping, increase stability, optimize Euclidean distance, and improve positioning accuracy. The main function of the WiFi fingerprint indoor positioning optimization algorithm fused with confidence is to reduce the error caused by real-time acquisition and set a lower confidence for the reference point that is reached with a small probability.

Description of drawings

1 is a schematic block diagram of a WiFi fingerprint indoor positioning method of fusion confidence proposed by the present invention;

FIG. 2 is a flowchart of offline collection of a WiFi fingerprint indoor positioning method of fusion confidence proposed by the present invention;

FIG. 3 is a flowchart of a WiFi fingerprint indoor positioning system with fusion confidence proposed by the present invention;

FIG. 4 is a schematic diagram of a WiFi fingerprint indoor positioning principle of fusion confidence level proposed by the present invention.

Detailed ways

The following describes in detail the embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary, only used to explain the present invention, and should not be construed as a limitation of the present invention.

These and other aspects of embodiments of the present invention will become apparent with reference to the following description and accompanying drawings. In these descriptions and drawings, some specific implementations of the embodiments of the invention are specifically disclosed to represent some ways of implementing the principles of the embodiments of the invention, but it should be understood that the scope of the embodiments of the invention is not limited by this limit. On the contrary, embodiments of the present invention include all changes, modifications and equivalents falling within the spirit and scope of the appended claims.

Inaccurate signal acquisition due to factors such as complex indoor environment, multipath effect, indoor fixed furniture and radio frequency interference will eventually lead to large online positioning errors. Aiming at this problem, the present invention proposes a WiFi fingerprint indoor positioning optimization algorithm fused with confidence, which can reduce jumps, increase stability, and improve positioning accuracy. The main function of the WiFi fingerprint indoor positioning optimization algorithm fused with confidence is to reduce the error caused by real-time acquisition and set a low confidence for the reference point that has a small probability of arrival. That is, when calculating the characteristic value of the positioning signal strength, it is no longer only the mean value, but the mean square error of the mean data set is also calculated and recorded, and the mean square error represents the quality of the signal. The mean square error is converted using the mean square error and confidence level mapping table, and finally uploaded to the server location fingerprint database together with the serial number of the reference point, the signal strength feature mean value set, and the geographic location coordinates.

One aspect of the embodiments of the present invention provides a WiFi fingerprint indoor positioning algorithm fused with confidence, including the following steps:

S101 , in the offline collection stage, use a collection device to collect multiple AP signal strength characteristic values used for positioning at each reference point, and preprocess the signal strength characteristic values, including calculating the mean value and the mean square error.

According to the formula A single AP at a certain reference point collects n times of signal strength characteristic values, and performs averaging processing to obtain an AP signal strength average value.

in, is the average signal strength of the AP, The sum of the signal strength eigenvalues collected n times for the AP.

In this embodiment, the collector uses a mobile smart phone to collect one by one in the same direction on the reference points that have been laid out. Each reference point is collected for 30 times, and each time is to collect the signal strength characteristic values of all APs, and the time is about one and a half minutes. through the formula The signal strength characteristic values of all APs collected at each reference point are respectively averaged to obtain the average value of the signal strength of each AP at each reference point.

in, is the mean value of the signal strength feature of the mth AP at a certain reference point, is the sum of the signal strength characteristic values of the mth AP of a certain reference point collected n times, n is the collection times of the collection device, i is an independent variable, m is the number of APs, A set of average signal strengths for all APs collected at a single reference point, respectively, A set of sums of signal strength eigenvalues is calculated for all APs collected at a single reference point.

After the mean value is performed, the mean square error processing can be performed according to the mean square error formula. Calculate the mean square error, is the characteristic value of the signal strength collected by a single AP for the nth time, is the average signal strength of the AP. The mean square error, also often called the standard deviation, is the arithmetic square root of the variance. Standard deviation can reflect the degree of dispersion of a data set. Two sets of data with the same mean may not have the same mean squared error. For example, the signal strength characteristic values of group X are (-95dBm, -85dBm, -75dBm, -65dBm, -55dBm, -45dBm), and the signal strength characteristic values of group B are (-73dBm, -72dBm, -71dBm, -69dBm, -68dBm, -67dBm). The average value of these two groups is -70dBm, but the mean square error of group A is about 17.08dBm, and the mean square error of group B is about 2.16dBm, indicating that the signal strength characteristic value of group B has a lower jump than the signal strength characteristic value of group A. , more stable, so the present invention uses the mean square error when converting.

It should be pointed out here that when the next reference point acquisition is performed, it should wait at this position for 10s to ensure the stability of the signal.

S102, the obtained mean square error of AP of all reference points is compared with a threshold; the threshold is the maximum mean square error in the virtual mapping table of mean square error interval and confidence level constructed by the present invention.

If the mean square error is greater than the threshold, the calculated mean value and the collected signal strength eigenvalues are subtracted one by one to obtain the absolute value, and then sorted in descending order, the signal strength eigenvalues corresponding to the first three largest absolute values are eliminated, and the remaining signal strength eigenvalues are recalculated. The mean squared error is recorded, and it is also judged whether it is within the threshold until it meets.

Among them, it should be noted that, if the signal of a certain AP is not very stable all the time, and the mean square error is always greater than the threshold, then re-sampling is performed.

If it is still greater than the threshold after re-sampling, the calculated mean value and the collected signal strength characteristic value are subtracted one by one to obtain the absolute value, and then sorted in descending order, and the signal strength characteristic value corresponding to the first three largest absolute values is eliminated. The mean square error of the intensity eigenvalues is recalculated and recorded until the last three signal intensity characteristic values are left, and the smallest mean square error and mean value in the whole process are recorded.

Wherein, the mean square error obtained for each AP at each reference point is converted to the mean square error according to the mean square error and confidence level mapping table to obtain the AP confidence level of the reference point.

The mean square error and confidence level mapping table is a virtual mapping table, which can be created in the program, and is divided into two types of elements: mean square error interval and confidence level.

As shown in the table below, each confidence level is mapped to a mean square error interval. When the mean square error of the signal strength characteristic value of an AP at a certain reference point falls within a mean square error interval in the following table, the AP The confidence of the signal strength characteristic value at the reference point is the corresponding confidence of the mean square error.

serial number mean square error interval Confidence 1 (10.00, 50.00) 0.60 2 (1.00, 10.00) 0.70 3 (0.50, 1.00) 0.80 4 (0.25, 0.50) 0.90 5 (0.00, 0.25) 1.00

Confidence is also called credibility. The threshold is determined as follows: there are 98 reference points and 4 APs in the entire experimental area, because each reference point needs to collect the signal strength characteristic values of 4 APs, that is, the mean square error. The number of calculations is 98*4=392, and the 392 mean square deviation values are eliminated by the outlier detection algorithm in the data analysis tool. The remaining mean square error interval is (0.00, 50.00), and the maximum mean square error is 50.00. as the threshold.

The method of determining the mean square error interval: first, with the same number as the standard, divide the mean square error interval 0.00 to 50.00 into 5 parts, that is, each mean square error interval can be preliminarily determined. After completion, the confidence level is decreased from 100% in 10% increments, and the mean square error interval is adjusted at the same time.

Next, continue to verify the mean square error interval based on the confidence level: if the mean square error interval is 0.00 to 0.25, find the reference point of the mean square error in this interval, re-collect the data and calculate the mean square error 30 times. In this interval, the reliability is 100%, that is, the confidence is 1.00. If not, the mean square error interval is enlarged until the confidence degree is 1.00. The same is true for other intervals, and then the virtual mapping between the mean square error interval and the confidence degree is determined. surface.

After the mean square error interval and the confidence level mapping table is determined, the mapping table can be used to determine the confidence level in other areas that have not been collected, without repeating the above process. If the mean square error calculated by a single AP is 0.10, the confidence level of the AP at the reference point is 1.00.

After the above processing, the confidence level of each AP of each reference point can be uploaded to the server location fingerprint database together with the geographic location coordinates of the reference point, the signal strength feature mean value set, and the serial number. The mean set of signal strength features in the database is recorded as: , N is the number of reference points, and m is the number of APs.

S103, use the confidence to optimize the Euclidean distance in the positioning matching stage.

The user uses the smart mobile device originally used for acquisition to collect the signal strength of a certain point to be located for a period of time. During this period of time, the signal strength characteristic values of all APs used for positioning will be collected, and the signal strength of all APs will be collected. The strength feature values are averaged respectively, and a set of AP signal strength feature mean values will be obtained, which is recorded as: , m is the number of APs, upload the mean set to the server, and add confidence in the Euclidean distance matching calculation with each group of feature data in the database, that is, multiply the relative value before the square and add in the calculation of the Euclidean distance. The reciprocal of the corresponding AP confidence, the geographic location coordinate corresponding to the reference point with the smallest Euclidean distance is used as the position estimate of the user's to-be-located point, and returned to the user, that is, the positioning is completed.

The original Euclidean distance formula is:

Let the confidence of the kth AP under the jth reference point in the database be , the Euclidean distance formula is changed to:

Where m is the number of APs, N is the total number of reference points in the database, A _k is the signal strength characteristic value of the k-th AP measured by the user at the point to be located, and B _jk is the k-th reference point in the database under the j-th reference point The signal strength characteristic value of each AP.

Another aspect of the embodiments of the present invention further provides a WiFi fingerprint indoor positioning system with integrated confidence, including a mobile terminal and a cloud server in network communication with the mobile terminal, wherein the mobile terminal includes a WiFi acquisition unit and a processing program, and the cloud The server includes a WiFi characteristic parameter storage unit and a positioning unit.

The WiFi acquisition unit and processing program are used to collect multiple AP signal strength characteristic values received by each reference point in the offline collection stage, and perform mean value and mean square error processing on the signal strength characteristic values and record, wherein each reference Points include multiple AP signal strength eigenvalues.

The WiFi characteristic parameter storage unit is used to store the MAC address, confidence level, geographic location coordinates of each AP, and the characteristic mean value and mean square error of each AP's signal strength at each reference point.

The positioning unit is used in the online positioning stage, and the confidence is added when the Euclidean distance matching calculation is performed between the signal strength characteristic value data of the user to be positioned point acquired in real time and each group of signal strength characteristic value data in the database, that is, when calculating the Euclidean distance. Multiplying the inverse of the confidence of the corresponding AP before the square sum and addition in the distance can achieve the effect of optimizing its distance.

The above-mentioned embodiment only represents an embodiment of the present invention, and its description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can also be made, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims (9)

1. a WiFi fingerprint indoor positioning algorithm of fusion confidence, is characterized in that: comprise the steps: In the offline collection stage, the collection device is used to collect multiple AP signal strength characteristic values used for positioning at each reference point, and the signal strength characteristic values are preprocessed to calculate the signal strength characteristic mean value and mean square error; After obtaining the mean square error, use the constructed virtual mapping table of mean square error interval and confidence level to determine the confidence level, and finally upload it to the server location fingerprint database together with the serial number of the reference point, the signal strength feature mean value set, and the geographic location coordinates; In the online positioning stage, the confidence is added when the Euclidean distance matching calculation is performed between the signal strength characteristic value data of the point to be located in real time and each group of characteristic data in the database, that is, before the square sum is added in the calculation of the Euclidean distance. The reciprocal of the corresponding AP confidence, the geographic location coordinate corresponding to the reference point with the smallest Euclidean distance is used as the position estimate of the user's to-be-located point, and returned to the user, that is, the positioning is completed. 2. The WiFi fingerprint indoor positioning algorithm of fusion confidence degree according to claim 1, is characterized in that: the step of described calculating signal strength characteristic mean value comprises: According to the formula A single AP at a certain reference point collects n times of signal strength characteristic values and performs averaging processing to obtain an AP signal strength average value; in, is the average signal strength of the AP, The sum of the signal strength eigenvalues collected n times for the AP; According to the formula The signal strength characteristic values of all APs collected at each reference point are respectively averaged to obtain the average value of the signal strength of each AP at each reference point; in, is the mean value of the signal strength feature of the mth AP at a certain reference point, is the sum of the signal strength characteristic values of the mth AP of a certain reference point collected n times, n is the collection times of the collection device, i is an independent variable, m is the number of APs, A set of average signal strengths for all APs collected at a single reference point, respectively, A set of sums of signal strength eigenvalues is obtained for all APs collected at a single reference point. 3. the WiFi fingerprint indoor positioning algorithm of fusion confidence degree according to claim 1, is characterized in that: the mean square error of described calculating signal strength characteristic value is: After each reference point receives the characteristic value of each AP signal strength, the mean value is processed and the mean square error is calculated at the same time; The mean square error formula is in, is the characteristic value of the signal strength collected by a single AP for the nth time, is the average signal strength of the AP. 4. The WiFi fingerprint indoor positioning algorithm of fusion confidence level according to claim 1 or 3, is characterized in that: described calculating signal strength characteristic value mean square error comprises the following steps: Each time the acquisition of the signal strength characteristic value of a reference point is completed, all APs calculate the mean square error separately and compare it with the threshold value. If the mean square error is greater than the threshold, the calculated mean value and the collected signal strength eigenvalues are subtracted one by one to obtain the absolute value, and then sorted in descending order, the signal strength eigenvalues corresponding to the first three largest absolute values are eliminated, and the remaining signal strength eigenvalues are recalculated. The mean squared error is recorded, and it is also judged whether it is within the threshold until it meets. 5. The WiFi fingerprint indoor positioning algorithm of fusion confidence according to claim 4, is characterized in that: whether described judgment mean square error is within the threshold value, also comprises later: If the signal of an AP is not very stable, the mean square error is always greater than the threshold, then re-sampling; If it is still greater than the threshold after re-sampling, the calculated mean value and the collected signal strength characteristic value are subtracted one by one to obtain the absolute value, and then sorted in descending order, and the signal strength characteristic value corresponding to the first three largest absolute values is eliminated. The mean square error of the intensity eigenvalues is recalculated and recorded until the last three signal intensity characteristic values are left, and the smallest mean square error and mean value in the whole process are recorded. 6. the WiFi fingerprint indoor positioning algorithm of fusion confidence degree according to claim 1, is characterized in that: the mean square error that each AP of each reference point obtains is carried out the conversion of mean square error according to mean square error and confidence degree mapping table to obtain this reference The AP confidence level of the point; the mean square error and confidence level mapping table is a virtual mapping table, which can be created in the program, and is divided into two types of elements: mean square error interval and confidence level. 7. The WiFi fingerprint indoor positioning algorithm of fusion confidence degree according to claim 1, is characterized in that: each AP confidence degree of each reference point that will obtain and the geographic location coordinates, sequence number, signal strength characteristic mean value set of this reference point They are uploaded to the server location fingerprint database together, and the mean set of signal strength features in the database is recorded as: , N is the number of reference points, and m is the number of APs. 8. The WiFi fingerprint indoor positioning algorithm of fusion confidence according to claim 1, is characterized in that: described in the online positioning stage, at a certain to-be-located point, first perform signal strength collection for a period of time, and in this period of time, Collect the signal strength feature values of all APs used for positioning, and calculate the mean value of the signal strength feature values of all APs respectively, and a set of AP signal strength feature mean values will be obtained, which is recorded as: , m is the number of APs, upload the mean set to the server, and add confidence in the Euclidean distance matching calculation with each group of feature data in the database, that is, multiply the relative value before the square and add in the calculation of the Euclidean distance. The reciprocal of the confidence of the corresponding AP, the geographic location coordinate corresponding to the reference point with the smallest Euclidean distance is used as the position estimate of the point to be located by the user, and returned to the user, that is, the positioning is completed; The original Euclidean distance formula is: Let the confidence of the kth AP under the jth reference point in the database be , the Euclidean distance formula is changed to: Where m is the number of APs, N is the total number of reference points in the database, A _k is the signal strength characteristic value of the k-th AP measured by the user at the point to be located, and B _jk is the k-th reference point in the database under the j-th reference point The signal strength characteristic value of each AP. 9. A WiFi fingerprint indoor positioning system with integrated confidence, characterized in that it comprises a mobile terminal, a cloud server for network communication with the mobile terminal, the mobile terminal includes a WiFi acquisition unit and a processing program, and the cloud server includes a WiFi feature Parameter storage unit and positioning unit; The WiFi acquisition unit and processing program are used to collect multiple AP signal strength characteristic values received by each reference point in the offline collection stage, and perform mean value and mean square error processing on the signal strength characteristic values and record, wherein each reference Points include multiple AP signal strength feature values; The WiFi characteristic parameter storage unit is used to store the MAC address, confidence level, geographic location coordinates of each AP, and the characteristic mean value and mean square error of the signal strength of each AP at each reference point; The positioning unit is used in the online positioning stage, and the confidence is added when the Euclidean distance matching calculation is performed between the signal strength characteristic value data of the user to be positioned point acquired in real time and each group of signal strength characteristic value data in the database, that is, when calculating the Euclidean distance. Multiply the reciprocal of the confidence of the corresponding AP before the square and sum in the distance, so as to achieve the effect of optimizing the distance.