CN113556699A - Sample set construction, indoor positioning model construction, indoor positioning method and device - Google Patents

Abstract

The invention discloses a method and device for sample set construction, indoor positioning model construction, and indoor positioning, and relates to the technical field of computers. A specific implementation of the method includes: acquiring multiple signal strengths of known reference points; determining a high probability occurrence interval according to the multiple signal strengths; selecting the signal strength belonging to the high probability occurrence interval from the multiple signal strengths; using For the signal strength belonging to the high probability occurrence interval, corresponding sample data is generated for the known reference points; the sample set is constructed by using the sample data corresponding to multiple known reference points. This embodiment can effectively remove dirty data in the sample set.

Description

Sample set construction, indoor positioning model construction, indoor positioning method and device

technical field

The invention relates to the field of computer technology, in particular to a method for constructing a sample set, a method for constructing an indoor positioning model, and an indoor positioning method and device.

Background technique

At present, an indoor positioning model is generally constructed by using a machine learning method based on the sample data in the sample set, and indoor positioning is performed according to the constructed indoor positioning model. For example, according to the constructed indoor positioning model, the robot running indoors is positioned or the driving route is planned.

In the process of realizing the present invention, the inventor found that there are at least the following problems in the prior art:

Due to the increasing complexity of the indoor environment and the increasing amount of dirty data in the sample set, how to effectively remove the dirty data in the sample set is a problem that needs to be solved.

SUMMARY OF THE INVENTION

In view of this, embodiments of the present invention provide a method and apparatus for sample set construction, indoor positioning model construction, and indoor positioning, which can effectively remove dirty data in the sample set.

To achieve the above object, according to an aspect of the embodiments of the present invention, a method for constructing a sample set is provided, including:

Obtain multiple signal strengths of known reference points;

determining a high probability occurrence interval according to the multiple signal strengths;

From the plurality of signal strengths, select a signal strength belonging to the high probability occurrence interval;

Generate corresponding sample data for the known reference point by using the signal strength belonging to the high probability occurrence interval;

Using a plurality of sample data corresponding to the known reference points, a sample set is constructed.

Preferably,

When the multiple signal strengths originate from at least two signal sources, and each of the signal sources corresponds to multiple signal strengths,

For a plurality of signal strengths corresponding to each of the signal sources, the step of determining a high probability occurrence interval is performed.

Preferably, generating corresponding sample data for the known reference point includes:

For a plurality of signal intensities originating from the same signal source and belonging to the high probability occurrence interval, execute: calculate the average value of the plurality of signal intensities, and use the average value as the sample data corresponding to the known reference point ;

or,

For the signal strength belonging to the high probability occurrence interval, perform: calculating the signal strength difference between every two of the signal sources, and using the signal strength difference as the sample data of the known reference point.

Preferably, the sample set construction method further includes:

From a plurality of the known reference points, select a plurality of target known reference points that satisfy a preset distance threshold for the unknown reference point;

assigning a corresponding weight to each of the target known reference points;

Using the weight and the sample data corresponding to the known reference point of the target, calculate the corresponding sample data for the unknown reference point;

The sample data corresponding to the unknown reference point is added to the sample set.

In a second aspect, an embodiment of the present invention provides a method for constructing an indoor positioning model, including:

obtaining a sample set, wherein the sample set is constructed from a plurality of sample data, and the sample data is generated from the signal strength belonging to the high probability occurrence interval;

Using the sample set, a support vector machine is iteratively trained to construct an indoor positioning model.

Preferably, the indoor positioning model construction method,

It further includes: setting a corresponding first value range for the parameters in the support vector machine;

Within the first value range, randomly assign initial parameters to the support vector machine;

Based on the initial parameters, the step of iteratively training the support vector machine is performed.

Preferably, the indoor positioning model construction method,

It further includes: setting an iterative step size for the parameters of the support vector machine;

The step of iteratively training the support vector machine includes:

Starting with the second iteration, using each iteration as the current iteration, execute:

determining the parameter corresponding to the previous iteration corresponding to the current iteration;

Increase or decrease the iteration step size for the parameter corresponding to the previous iteration, as the parameter of the current iteration;

training the support vector machine based on the parameters of the current iteration;

When the training result satisfies the preset first termination condition, the iteration is terminated.

Preferably, the indoor positioning model construction method,

It further includes: setting a corresponding second value range for the parameter in the support vector machine;

Initialize the firework population within the second value range;

The step of iteratively training the support vector machine includes:

For each iteration, execute:

determining a firework population, and determining the explosion sparks and variant sparks of the fireworks in the firework population;

calculating the fitness of the explosion spark and the mutated spark;

Determine whether the result of the current iteration satisfies the second termination condition,

If yes, take the explosion spark with the least fitness or the mutation spark with the least fitness as the training result, and end the current process;

Otherwise, from the explosion sparks and the variation sparks, multiple target sparks are selected to form a firework population, wherein the target sparks are used as fireworks in the firework population.

In a third aspect, an embodiment of the present invention provides an indoor positioning method implemented based on an indoor positioning model constructed by any of the above methods, including:

Obtain the signal fingerprint information of the to-be-located point;

The signal fingerprint information is input into the indoor positioning model to obtain the position information of the to-be-located point.

Preferably, the acquiring the signal fingerprint information of the point to be located includes:

The collected signal strength of the to-be-located point is used as the signal fingerprint information;

or,

For the collected signal strengths of at least two signal sources at the to-be-located point, calculate the signal strength difference corresponding to each of the two signal sources, and use the signal strength difference as the signal fingerprint information.

In a fourth aspect, an embodiment of the present invention provides an apparatus for constructing a sample set, including:

a signal acquisition unit for acquiring multiple signal strengths of known reference points;

a signal processing unit, configured to determine a high probability occurrence interval according to a plurality of signal strengths obtained by the signal acquisition unit; and select a signal strength belonging to the high probability occurrence interval from the plurality of signal strengths;

a sample set construction unit, configured to generate corresponding sample data for the known reference points by using the signal strengths selected by the signal processing unit and belonging to the high probability occurrence interval; using a plurality of the known reference points corresponding to sample data to construct a sample set.

Preferably,

The signal processing unit is configured to, when the multiple signal strengths originate from at least two signal sources, and each of the signal sources corresponds to multiple signal strengths, for each of the signal sources corresponding to multiple signal strengths , and perform the step of determining a high probability occurrence interval according to the multiple signal strengths.

Preferably,

The signal processing unit is further configured to select a plurality of target known reference points that satisfy a preset distance threshold for the unknown reference point from a plurality of the known reference points; for each of the target known reference points assign corresponding weights;

The sample set construction unit is further configured to use the weight and the sample data corresponding to the target known reference point to calculate the corresponding sample data for the unknown reference point; add the sample data corresponding to the unknown reference point into the sample set.

In a fifth aspect, an embodiment of the present invention provides an apparatus for constructing an indoor positioning model, including:

a sample set obtaining unit, configured to obtain a sample set, wherein the sample set is constructed from a plurality of sample data, and the sample data is generated from the signal intensity belonging to the high probability occurrence interval;

The model building unit is configured to use the sample set obtained by the sample set obtaining unit to iteratively train a support vector machine to construct an indoor positioning model.

In a sixth aspect, an embodiment of the present invention provides an indoor positioning device, including:

an information acquisition unit, used for acquiring the signal fingerprint information of the to-be-located point;

A position determination unit, configured to input the signal fingerprint information acquired by the information acquisition unit into the indoor positioning model to obtain the position information of the to-be-located point.

An embodiment of the above invention has the following advantages or beneficial effects: for the acquired data, such as signal strength, most of them are normal data, and the distribution of normal data is generally concentrated. range or interval, and a small part of abnormal data, that is, dirty data, is often outside the range or interval where normal data is located. Based on this, the solution provided by the embodiment of the present invention determines a high probability occurrence interval according to multiple signal strengths; using The signal strength belonging to the high probability occurrence interval generates corresponding sample data for the known reference points; the sample data corresponding to multiple known reference points is used to construct a sample set, and the high probability occurrence interval is used to realize the removal of dirty data, that is, the construction of the sample The sample data of the set is obtained from the data after removing the dirty data, which can effectively remove the dirty data in the sample set.

Further effects of the above non-conventional alternatives will be described below in conjunction with specific embodiments.

Description of drawings

The accompanying drawings are used for better understanding of the present invention and do not constitute an improper limitation of the present invention. in:

1 is a schematic diagram of an application scenario according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a main flow of a method for constructing a sample set according to an embodiment of the present invention;

3 is a schematic diagram of a main flow of a method for constructing a sample set according to another embodiment of the present invention;

4 is a schematic diagram of a sample data collection site according to an embodiment of the present invention;

5 is a schematic diagram of a main process of an indoor positioning model construction method according to an embodiment of the present invention;

6 is a schematic diagram of a main process of iteratively training a support vector machine according to an embodiment of the present invention;

7 is a schematic diagram of a main process of iteratively training a support vector machine according to another embodiment of the present invention;

8 is a schematic diagram of a main flow of an indoor positioning method according to an embodiment of the present invention;

9 is a schematic diagram of main units of a sample set construction apparatus according to an embodiment of the present invention;

10 is a schematic diagram of the main units of an indoor positioning model building apparatus according to an embodiment of the present invention;

11 is a schematic diagram of the main unit of an indoor positioning device according to an embodiment of the present invention;

FIG. 12 is an exemplary system architecture diagram to which an embodiment of the present invention may be applied;

FIG. 13 is a schematic structural diagram of a computer system suitable for implementing a terminal device or a server according to an embodiment of the present invention.

Detailed ways

Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, which include various details of the embodiments of the present invention to facilitate understanding and should be considered as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted from the following description for clarity and conciseness.

An application scenario of the solution provided by the embodiment of the present invention may be shown in FIG. 1 . In Fig. 1, since the signal strengths generated by signal generating devices such as wireless routers and communication towers that can receive signals at different locations are different, a model of the relationship between signal strength and location can be constructed based on the signal strengths. Then, after acquiring the signal strength of an unknown location, the unknown location can be located. Through signal acquisition equipment such as mobile phones, computers, tablets, etc., the signal strengths sent by the signal generating equipment (AP1, AP2, . The obtained signal strength can be stored in the corresponding database. The storage of the location information and the collected signal strength may be as shown in Table 1 below.

Table 1

reference point location information signal strength I₁ (x₁, y₁) (RSSI_11AP1, RSSI_12AP1, ..., RSSI_1ab) I₂ (x₂, y₂) (RSSI_21AP1, RSSI_22AP1, ..., RSSI_2ab) ... ... ... I_j (x_j, y_j) (RSSI_j1AP1, RSSI_j2AP1, ..., RSSI_jab)

Among them, I _j represents the jth known reference point; RSSI _jab represents the signal strength of the signal source b collected at the jth known reference point a for the ath time; the value of j is a positive integer not less than 1; The value of a is a positive integer not less than 1; the value range of b is AP1, AP2, ..., APn; (x _j , y _j ) represents the position information (position information) corresponding to the jth known reference point coordinate).

The position coordinates of the known reference points are determined based on the same coordinate system, for example, the coordinate system is a coordinate system arbitrarily constructed indoors.

Based on the collected signal strengths of the known reference points, a sample set is constructed, and a data model is trained based on the sample set. The position of the unknown point can be obtained according to the signal strength of the unknown point based on the trained data model. It is used in indoor location determination, management and path planning, for example, determining the indoor location of indoor sweeping robots, food delivery robots, warehouse handling equipment, etc., and planning driving paths for indoor cleaning robots, food delivery robots, warehouse handling equipment, etc.

A known reference point means that the location information has been determined and the corresponding signal strength has been collected.

Correspondingly, the unknown reference point means that the location information has been determined, but the corresponding signal strength has not been collected.

Wherein, the location information may be a coordinate position, and for a known reference point and/or an unknown reference point with a relatively small distribution area or distributed indoors, the coordinate position may be a pre-built coordinate system (for example, a user based on the characteristics of the building structure) The coordinate position in the constructed coordinate system, etc.); for a known reference point and/or an unknown reference point with a relatively large distribution area, the coordinate position may also be longitude and latitude.

FIG. 2 is a method for constructing a sample set according to an embodiment of the present invention. As shown in Figure 1, the sample set construction method may include the following steps:

201: Acquire multiple signal strengths of known reference points;

202: Determine a high probability occurrence interval according to multiple signal strengths;

203: Select a signal strength belonging to a high probability occurrence interval from a plurality of signal strengths;

204: Generate corresponding sample data for the known reference point by using the signal strength belonging to the high probability occurrence interval;

205: Construct a sample set by using sample data corresponding to a plurality of known reference points.

For the acquired data such as signal strength, most of them are normal data, and the distribution of normal data is generally concentrated. In general, the range or interval where normal data is located can be defined, while a small part of abnormal data is dirty data. It is often outside the range or interval where the normal data is located. Based on this, the solution provided by the embodiment of the present invention determines the high probability occurrence interval according to multiple signal strengths; uses the signal strength belonging to the high probability occurrence interval as the known reference point. Generate corresponding sample data; use the sample data corresponding to multiple known reference points to construct a sample set, and use a high probability occurrence interval to remove dirty data, that is, the sample data for constructing a sample set is obtained from the data after removing the dirty data , which can effectively remove the dirty data in the sample set.

The specific implementation of step 201 may include: receiving the signal strength collected by the signal collecting device at a known reference point. The signal acquisition device collects the signal strength at a known reference point. For example, for the WiFi signal, the WiFi signal acquisition software WiFiScan is installed on the laptop to perform the acquisition, and the above step 201 is implemented.

The specific implementation of step 202 may include: determining the high probability occurrence interval as the interval range obtained by the following calculation formula (1).

Calculation formula (1)

Among them, n _j represents the number of signal strengths corresponding to the jth known reference point; RSSI _ji represents the ith signal strength corresponding to the jth known reference point; j and i are both positive integers not less than 1.

Correspondingly, among multiple signal strengths, greater than or equal to

且小于等于

and less than or equal to

的信号强度，即为属于高概率发生区间的信号强度。

The signal strength of , that is, the signal strength belonging to the high probability occurrence interval.

In an embodiment of the present invention, when multiple signal strengths originate from at least two signal sources, and each signal source corresponds to multiple signal strengths, for the multiple signal strengths corresponding to each signal source, the Intensity, steps to determine the high probability interval.

For example, the signal strength obtained at the known reference point A comes from the signal sources AP1, AP2 and AP3. in,

_The signal strengths corresponding to the signal source _AP1 are RSSI ₁ , RSSI ₂ , _{RSSI 3} , . RSSI ₃ , ..., RSSI _N determine the high probability occurrence interval of the signal source AP1, and select the signal strength belonging to the high probability occurrence interval of the signal source AP1 from RSSI ₁ , RSSI ₂ , RSSI ₃ , ..., RSSI _N , so as to be the signal strength of the signal source AP1 Knowing that the reference point A generates the sample data corresponding to the signal source AP1;

The signal strengths corresponding to the signal source AP2 are RSSI _{N +1} , RSSI _N+2 , RSSI _{N +3} , . _1. RSSI _N+2 , RSSI _N+3 ,..., RSSI _2N determine the high probability occurrence interval of the signal source AP2, and select from RSSI _N+1 , RSSI _N+2 , RSSI _N+3 ,..., RSSI _2N belonging to The signal strength of the high probability occurrence interval of the signal source AP2, so as to generate the sample data corresponding to the signal source AP2 for the known reference point A;

The signal strength corresponding to the signal source AP3 is:

RSSI _2N+1 , RSSI _2N+2 , RSSI _2N+3 , …, RSSI _3N ; then pass

RSSI _2N+1 , RSSI _2N+2 , RSSI _2N+3 , …, RSSI _3N determine the high probability occurrence interval of the signal source AP3, and from RSSI _2N+1 , RSSI _2N+2 , _{RSSI 2N+3} , … Select the signal strength belonging to the high probability occurrence interval of the signal source AP3, so as to generate the sample data corresponding to the signal source AP3 for the known reference point A.

Through the above embodiment, the signal strength is calculated according to the known reference point and the signal source, and the error caused by the difference in the distance from different signal sources to the known reference point and the difference in signal strength generated by different signal sources is avoided.

In an embodiment of the present invention, there may be two specific implementation manners for generating the sample data in the foregoing step 204 .

The specific implementation of the first sample data:

For a plurality of signal intensities originating from the same signal source and belonging to a high probability occurrence interval, execute: calculate the average value of the plurality of signal intensities, and use the average value as the sample data corresponding to the known reference point.

Wherein, the following calculation formula (2) can be used to calculate the average value of multiple signal strengths;

Calculation formula (2):

表征信号源APn所对应的平均值；m_APn表征源于信号源APn且属于高概率发生区间的信号强度的个数；RSSI_tAPn表征来源于信号源APn且属于高概率发生区间的第t个信号强度，t为不大于m的正整数，m为不小于2的正整数。in,

Represents the average value corresponding to the signal source APn; m _APn represents the number of signal strengths originating from the signal source APn and belonging to the high probability occurrence interval; RSSI _tAPn represents the t-th signal originating from the signal source APn and belonging to the high probability occurrence interval Intensity, t is a positive integer not greater than m, and m is a positive integer not less than 2.

The specific implementation of the second sample data:

For the signal strength belonging to the high probability occurrence interval, execute: calculate the signal strength difference between every two signal sources, and use the signal strength difference as the sample data of the known reference point.

Among them, the signal strength difference between each two signal sources can be calculated by using the following calculation formula (3):

Calculation formula (3):

表征由源于信号源APi且属于高概率发生区间的多个信号强度计算得到的平均值；

表征由源于信号源APj且属于高概率发生区间的多个信号强度计算得到的平均值。Among them, RSSU APi _-APj represents the signal strength difference between the signal source APi and the signal source APj;

Characterize the average value calculated from multiple signal strengths originating from the signal source APi and belonging to the high probability occurrence interval;

Characterizes the mean value calculated from multiple signal strengths originating from the signal source APj and belonging to the high probability occurrence interval.

In an embodiment of the present invention, as shown in FIG. 3 , the above-mentioned method for constructing a sample set may further include the following steps:

301: From a plurality of known reference points, select a plurality of target known reference points that satisfy a preset distance threshold for the unknown reference point;

302: Assign a corresponding weight to each target known reference point;

303: Calculate the corresponding sample data for the unknown reference point by using the weight and the sample data corresponding to the target known reference point;

304: Add sample data corresponding to the unknown reference point to the sample set.

Through the above process, the sample data is added to the sample set to make the sample set more abundant. Since the sample data of the unknown reference point is obtained from the sample data of the known reference point, the steps of acquiring multiple signal strengths of the unknown reference point and determining the high probability occurrence interval are avoided. The intensity process generally takes a long time, and the sample data based on the known reference point can effectively shorten the time required to obtain the sample data of the unknown reference point. Therefore, the solution provided in this embodiment can effectively improve the construction of the sample data. set efficiency.

Generally, known reference points and unknown reference points exist alternately. The sample data collection site is given in Figure 4, including a rectangular area of 4 rooms and a corridor part. A grid of 1 m × 1 m is divided for the rectangular area, and the points in each grid (such as the center point or non-center point of the grid) are known reference points or unknown reference points, among which, Figure 4 shows The output C1, C2, C3, C4, C5, C6, C7, C8, C9, C10 are some known reference points, the corresponding c1, c2, c3, c4, c5, c6, c7, c8, c9, c10 , c11 is a partial unknown reference point. Among them, the distribution of some known reference points and some unknown reference points given in Figure 4 is only an example, and other distribution methods are also possible (it only needs to ensure that there is a known know the reference point).

Meeting the preset distance threshold generally refers to a known reference point whose distance from the unknown reference point is not more than 2 times the grid side length (for example, if the grid side length is 1 meter, then satisfying the preset distance threshold is the distance between the unknown reference point and the unknown reference point). Known reference point greater than 2 meters) to ensure that the selected target known reference point is basically adjacent to the unknown reference point.

Wherein, step 302 and step 303 can be calculated by adopting the kriging interpolation method. Kriging interpolation method is to perform optimal unbiased estimation of unknown points by referring to the characteristics of known points. First, a variogram model corresponding to the reference point information in the positioning area is constructed, and then on the basis of this structural analysis Solve the kriging equation to determine, obtain the coordinate position of the unknown estimated point, and obtain the characteristic attribute value of the point to be estimated through cross-validation.

Among them, the corresponding weight assigned to each target known reference point can be calculated by the following calculation formula group:

Calculation formula group:

Among them, λ _i represents the weight corresponding to the i-th target known parameter point selected in step 301; h represents the total number of target known parameter points selected in step 301; Cov(x _i , x _j ) represents The covariance between the signal strength corresponding to the selected i-th target known parameter point and the signal strength corresponding to the selected j-th target known parameter point (the j-th target known parameter point is indicated as dividing the i-th target known parameter point Any one of the target known parameter data points randomly selected outside the target known parameter points); Cov(x ₀ , x _i ) represents the signal strength corresponding to the unknown reference point and the signal corresponding to the i-th target known parameter point Covariance between intensities; μ characterizes Lagrangian coefficients.

The specific process obtained by the above calculation formula group:

Assuming that the regionalized variable Z(x) (the variable whose signal strength is in the region) satisfies the second-order stationary in the entire study area,

That is, assuming that the mathematical expectation of Z(x) exists and is equal to a constant: E[Z(x)]=m(constant), the covariance Cov(x _i , x _j ) of Z(x) exists and is only related to the difference between the two points relative position between them.

Or assume that the regionalized variable Z(x) satisfies the eigen hypothesis: E[Z(x _i )-Z(x _j )]=0, the variance of the increments exists and is stationary: Var[Z(x _i )-Z(x _j )]=E[Z(x _i )-Z(x _j )] ² .

According to the requirement of unbiasedness: E[Z ^* (x ₀ )]=E[Z(x ₀ )], it can be deduced from the above:

The estimated variance is minimized under unbiased conditions, namely:

where μ is the Lagrangian coefficient, so the above calculation formula group can be obtained, so as to obtain the weight λ _{i (i=1, 2,..., h} )° of the known reference points of each target

Based on the weights calculated above, the following calculation formula (4) is used to calculate the sample data corresponding to the unknown reference points.

Calculation formula (4):

Among them, Z ^* (x ₀ ) represents the signal strength of the unknown reference point x ₀ ; Z(x _i ) represents the signal strength of the i-th target known parameter point; λ _i represents the i-th target selected in step S301 has been weights corresponding to known parameter points; h represents the total number of target known parameter points selected in step S301.

That is, after obtaining the weights λ _{i (i=1, 2, . . . , h)} , the signal strength Z ^* (x ₀ ) of the unknown reference point x ₀ can be obtained.

In addition, in the above-mentioned sample set construction process, two types of acquisition devices can be selected to collect multiple signal intensities at known reference points, so that the multiple signal intensities collected by one type of acquisition device can be used as the training sample set. The multiple signal strengths collected by another type of collection device are used as the test sample set. Each type of acquisition device collects signal strength at each known reference point for no less than 12 times.

For example, for the sample data collection site shown in Figure 4, a type of PC-side WiFi signal collection software WiFiScan is used to collect data for each known reference point 12 times, once per second, and finally a training sample set is obtained. The sample set for testing the positioning accuracy (the sample set used in the model testing phase) uses the WiFi signal acquisition software of another model of notebook randomly in each area in Figure 4 (room 1, room 2, room 3, room 4 and Corridor) to collect multiple signal strengths to construct a test sample set.

As shown in FIG. 5 , an embodiment of the present invention provides a method for constructing an indoor positioning model, and the method for constructing an indoor positioning model may include the following steps:

501: Obtain a sample set, where the sample set is constructed from multiple sample data, and the sample data is generated from signal strengths belonging to a high probability occurrence interval;

502: Use the sample set to iteratively train the support vector machine to construct an indoor positioning model.

Since the sample data in the sample set is generated by the signal intensity belonging to the high probability occurrence interval, the validity and accuracy of the sample data in the entire sample set are greatly improved, and the premise that the validity and accuracy of the sample data in the sample set are greatly improved The accuracy of the indoor positioning model constructed based on this sample set has also been effectively improved.

In an embodiment of the present invention, the above-mentioned method for constructing an indoor positioning model may further include: setting a corresponding first value range for the parameters in the support vector machine; within the first value range, randomly assigning initial parameters to the support vector machine ; Based on the initial parameters, perform the steps of iteratively training the support vector machine. The first value ranges from 0.1 to 10.

By setting the corresponding first value range for the parameters in the support vector machine above; within the first value range, randomly assigning initial parameters to the support vector machine, effectively reducing the parameter range, and ensuring the accuracy of the training results. At the same time, the number of iterations can be effectively reduced.

In an embodiment of the present invention, the above-mentioned method for constructing an indoor positioning model may further include: setting an iterative step size for the parameters of the support vector machine; correspondingly, as shown in FIG. 6 , starting from the second iteration, each iteration is taken as For the current iteration, steps 601 to 604 can be performed:

601: Determine the parameters corresponding to the previous iteration corresponding to the current iteration;

602: Increase or decrease the iteration step size for the parameter corresponding to the previous iteration, as the parameter of the current iteration;

603: Train the support vector machine based on the parameters of the current iteration;

604: Terminate the iteration when the training result satisfies the preset first termination condition.

The specific implementation manner of setting the iteration step size for the parameters of the support vector machine above is as follows: the previous iteration parameter of each iteration is multiplied by 0.1 or 10 as a step size, so as to obtain the current iteration parameter. For example, a parameter of the previous iteration is 0.1, a step size of the current iteration is 0.1×0.1=0.01, and the parameter is 0.1+0.01=0.11; for another example, a parameter of the previous iteration is 10, a step of the current iteration The length is 10×0.1=1, and the parameter is 10-1=9.

The first termination condition may be: the number of iterations reaches a preset threshold of the first iteration number; or, the first termination condition may be: the difference between the result of the current iteration and the result of the previous iteration is not greater than the preset threshold Alternatively, the above-mentioned first termination condition may be: the difference between the positioning result generated by indoor positioning of the model generated by the current iteration and the actual position is not greater than a preset difference threshold.

It is worth noting that, after the parameter value is determined in the embodiment shown in FIG. 6 , the parameter value can be refined and searched for a more accurate value. There are two specific implementations for the refinement of the parameter value to search for a more accurate value. The first method is to reduce the step size to take the value, further determine an interval for the determined parameter value, and use the reduced step in the interval. The iterative process shown in FIG. 6 above is further performed. The second way is: using the iterative process shown in FIG. 7 to implement.

It should be noted that the specific process given in FIG. 7 below can be used as a parallel embodiment of the embodiment shown in FIG. 6 , and can also be used as a further limitation of the embodiment shown in FIG. 6 .

In an embodiment of the present invention, for iterative training of a support vector machine, as shown in FIG. 7 , the above-mentioned method for constructing an indoor positioning model may include:

701: Set a corresponding second value range for a parameter in the support vector machine;

When the specific process shown in FIG. 7 is a parallel embodiment of the embodiment shown in FIG. 6 , the second value range may be consistent with the first value range, or may be different from the first value range. The value range can be set manually. When the specific process shown in FIG. 7 is further defined as the embodiment shown in FIG. 6 , the second value range is a value determined based on the parameters generated by the last iteration obtained in the embodiment shown in FIG. 6 . interval.

702: Initialize the firework population within the second value range;

For each iteration, step 703 to step 707 are performed:

703: Determine the firework population, and determine the explosion sparks and variation sparks of the fireworks in the firework population;

704: Calculate the fitness of explosion sparks and mutant sparks;

705: Determine whether the result of the current iteration satisfies the second termination condition, if yes, execute step 706; otherwise, execute step 707;

706: Use the explosion spark with the least fitness or the mutation spark with the minimum fitness as the training result, and end the current process;

707: From the explosion sparks and the variation sparks, select multiple target sparks to form a firework population, wherein the target sparks are used as fireworks in the firework population.

Wherein, in the above step 702, the number of sparks S _i and the explosion amplitude radius A _i generated by each explosion of the fireworks g _i in the fireworks population can be calculated by calculating the explosion operator;

Among them, the number of sparks Si can be calculated by the following calculation formula (5) _{, and the explosion range radius A i can} be calculated by the following calculation formula (6).

Calculation formula (5):

Calculation formula (6):

表征最大的爆炸幅度，ε表征一极小常数以免出现除零操作；g_i表征当前烟花种群中第i个爆炸烟花；p表征当前烟花种群中包括的爆炸烟花总数。where w represents a constant to control the total number of sparks, Y _max =max(f( _gi )) represents the maximum fitness in the current firework population, Y _min =min(f( _gi )) represents the minimum fitness in the current population value,

Represents the maximum explosion amplitude, ε represents a minimal constant to avoid division by zero; g _i represents the ith exploding firework in the current firework population; p represents the total number of exploding fireworks included in the current firework population.

In addition, f( _gi )=1-D( _gi ), where f( _gi ) represents the fitness corresponding to the fireworks _gi in the current population; D( _gi ) represents the location corresponding to the fireworks _gi in the current population Precision accuracy.

In addition, to determine the mutation spark, the Gaussian mutation spark can be calculated by the mutation operator. Specifically: randomly select a spark g _i ^u and perform a Gaussian mutation operation on the u-th dimension, where v is a Gaussian random value obeying v～N(1, 1).

Mutation operator:

g _i ^u = g _i ^u ×v

The out-of-bound sparks are mapped into the feasible region using the modulo operation mapping rule.

Model operation mapping rules:

G _i ^u =g _min ^u +|G _i ^u |%(g _max ^u -g _min ^u )

Among them, G _i ^u represents the position of the out-of-bounds i-th spark on the u-th dimension; g _max ^u and g _min ^u represent the upper and lower boundaries on the u-th dimension, respectively.

The specific implementation of the above-mentioned step 704 is that the fitness of the explosive spark or the mutated spark can be obtained by subtracting the accuracy of the corresponding positioning accuracy of the explosive firework or the mutated spark by 1.

The second termination condition may be that the number of iterations reaches a preset second threshold of the number of iterations.

The specific implementation manner of the above step 707 may be:

The elite strategy is used to select the spark with the smallest fitness function value in this iteration (ie the optimal spark) as the explosion fireworks of the next iteration, and the remaining N fireworks of the next iteration are selected by roulette gambling. The probability formula p(G _i ) for the spark G _i to be selected is as follows.

where R(G _i ) is the Euclidean distance between two individuals.

R(G _i )=∑ _j∈K ||G _i -G _j ||

Among them, G _i represents the ith spark among the explosion sparks and mutation sparks generated by the current iteration; G _j represents any one of the explosion sparks and mutation sparks generated by the current iteration; K represents the explosion fireworks to which G _i and G _j belong The resulting collection of sparks;

According to the order of probability from large to small, N sparks are selected to form the firework population of the next iteration.

As shown in FIG. 8 , an embodiment of the present invention provides an indoor positioning method implemented based on the indoor positioning model constructed in the foregoing embodiment. The indoor positioning method may include the following steps:

801: Obtain the signal fingerprint information of the point to be located;

802: Input the signal fingerprint information into the indoor positioning model to obtain the position information of the point to be positioned.

The above step 801 may be: collecting the signal strength of the to-be-located point, and using the collected signal strength as the signal fingerprint information;

The above step 801 may also be: collecting the signal strengths of at least two signal sources at the to-be-located point, calculating the signal strength difference corresponding to each two signal sources, and using the signal strength difference as the signal fingerprint information.

Using the signal strength difference as the signal fingerprint information can remove the error caused by the hardware of the device that collects the signal strength, thereby effectively improving the positioning accuracy.

As shown in FIG. 9 , an embodiment of the present invention provides an apparatus 900 for constructing a sample set, and the apparatus 900 for constructing a sample set may include:

a signal acquisition unit 901, configured to acquire multiple signal strengths of known reference points;

The signal processing unit 902 is configured to determine the high probability occurrence interval according to the multiple signal strengths obtained by the signal obtaining unit 901; from the multiple signal strengths, select the signal strength belonging to the high probability occurrence interval;

The sample set construction unit 903 is used to generate corresponding sample data for the known reference points by using the signal strengths belonging to the high probability occurrence interval selected by the signal processing unit 902; set.

In an embodiment of the present invention, the signal processing unit 902 is configured to, when multiple signal strengths originate from at least two signal sources, and each signal source corresponds to multiple signal strengths, for multiple signals corresponding to each signal source Intensity, perform the steps to determine the high probability occurrence interval.

In an embodiment of the present invention, the sample set construction unit 903 is configured to, for multiple signal intensities originating from the same signal source and belonging to the high probability occurrence interval, perform: calculating the average value of the multiple signal intensities, and taking the average value as the Sample data corresponding to known reference points.

In an embodiment of the present invention, the sample set construction unit 903 is configured to, for the signal strength belonging to the high probability occurrence interval, perform: calculate the signal strength difference between every two signal sources, and use the signal strength difference as a known reference point sample data.

In an embodiment of the present invention, the signal processing unit 902 is further configured to select a plurality of target known reference points that satisfy a preset distance threshold for the unknown reference point from a plurality of known reference points; Know the reference point to assign the corresponding weight;

The sample set construction unit 903 is further configured to calculate the corresponding sample data for the unknown reference point by using the weight assigned by the signal processing unit 902 and the sample data corresponding to the target known reference point; adding the sample data corresponding to the unknown reference point into the sample set .

As shown in FIG. 10 , an embodiment of the present invention provides an apparatus 1000 for constructing an indoor positioning model. The apparatus 1000 for constructing an indoor positioning model may include:

A sample set obtaining unit 1001, configured to obtain a sample set, wherein the sample set is constructed by a plurality of sample data, and the sample data is generated by the signal strength belonging to the high probability occurrence interval;

The model building unit 1002 is configured to use the sample set obtained by the sample set obtaining unit to iteratively train the support vector machine to construct an indoor positioning model.

Wherein, the sample set obtaining unit 1001 may obtain the sample set constructed by the sample set constructing apparatus 900 of the above-mentioned embodiment shown in FIG. 9 .

In an embodiment of the present invention, the model building unit 1002 is further configured to set a corresponding first value range for the parameters in the support vector machine; within the first value range, randomly assign initial parameters to the support vector machine; Based on the initial parameters, the steps of iteratively training the support vector machine are performed.

In an embodiment of the present invention, the model building unit 1002 is further configured to set an iteration step size for the parameters of the support vector machine; starting from the second iteration, take each iteration as the current iteration, and execute: determine the corresponding value of the current iteration The parameters corresponding to the previous iteration; increase or decrease the iteration step size for the parameters corresponding to the previous iteration, as the parameters of the current iteration; train the support vector machine based on the parameters of the current iteration; when the training result satisfies the preset first termination condition , the iteration is terminated.

In an embodiment of the present invention, the model building unit 1002 is further configured to set a corresponding second value range for the parameters in the support vector machine; initialize the firework population within the second value range; for each iteration , execute: determine the fireworks population, determine the explosion sparks and mutation sparks of the fireworks in the fireworks population; calculate the fitness of the explosion sparks and mutation sparks; determine whether the result of the current iteration satisfies the second termination condition, if so, set the fitness to the smallest The explosion spark or the mutation spark with the least fitness is used as the training result, and the current process is ended; otherwise, multiple target sparks are selected from the explosion sparks and mutation sparks to form a firework population, and the target spark is used as the fireworks in the firework population.

As shown in FIG. 11 , an embodiment of the present invention provides an indoor positioning apparatus 1100, and the indoor positioning apparatus 1100 may include:

An information acquisition unit 1101, configured to acquire the signal fingerprint information of the point to be located;

The position determining unit 1102 is configured to input the signal fingerprint information obtained by the information obtaining unit into the indoor positioning model to obtain the position information of the to-be-located point.

In an embodiment of the present invention, the information acquisition unit 1101 is configured to use the collected signal strength of the point to be located as the signal fingerprint information.

In an embodiment of the present invention, the information acquisition unit 1101 is further configured to calculate the signal strength difference corresponding to each two signal sources according to the collected signal strengths of at least two signal sources at the to-be-located point, and use the signal strength difference as Signal fingerprint information.

It should be noted that the indoor positioning model used by the position determining unit 1102 may be derived from the indoor positioning model constructed by the indoor positioning model constructing apparatus 1000 provided in the above embodiment.

In addition, the above-mentioned sample set construction apparatus 900, indoor positioning model construction apparatus 1000, and indoor positioning apparatus 1100 may be integrated into the same device or system in the form of plug-ins.

In addition, the above-mentioned sample set construction apparatus 900 , indoor positioning model construction apparatus 1000 , and indoor positioning apparatus 1100 may be integrated into a device or device that integrates sample construction, indoor positioning model construction, and indoor positioning.

In addition, the above-mentioned sample set construction apparatus 900 and indoor positioning model construction apparatus 1000 may be integrated into the same device, and the indoor positioning apparatus 1100 may be in another device; the sample set construction apparatus 900 may be in a separate device, and the indoor positioning model construction apparatus 1000 and the indoor positioning apparatus The 1100 can be integrated into the same device.

12 shows an exemplary system architecture 1200 of a sample set construction method or a sample set construction apparatus or an indoor positioning model construction method or an indoor positioning model construction apparatus or an indoor positioning method or an indoor positioning apparatus to which embodiments of the present invention can be applied.

As shown in FIG. 12 , the system architecture 1200 may include terminal devices 1201 , 1202 , 1203 , a network 1204 , a server 1205 , a database 1206 and an automated mobile device 1207 . The network 1204 is used between the terminal devices 1201, 1202, 1203 and the server 1205, between the terminal devices 1201, 1202, 1203 and the database 1206, between the server 1205 and the database 1206, between the database 1206 and the automatic mobile device 1207, between the server 1205 and the automated mobile device 1207 provide the medium of the communication link. Network 1204 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user can use the terminal devices 1201, 1202, 1203 to interact with the server 1205 through the network 1204 to send or receive information and the like. The terminal devices 1201, 1202, 1203 interact with the database 1206 through the network 1204 to send or receive information related to signal strength and the like. The database 1206 interacts with the server 1205 to send sample sets or receive constructed indoor positioning models, and the like. The automated mobile device 1207 interacts with the server 1205 to send signal fingerprint information or receive positioning results. Various communication client applications can be installed on the terminal devices 1201, 1202, 1203 and the automatic mobile device 1207, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. ( example only).

The terminal devices 1201, 1202, 1203 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

The automatic mobile device 1207 may be an automatic warehouse handling device, a handling robot, a food delivery robot, a cleaning robot, and the like.

The server 1205 may be a server that provides various services, such as a background management server (just an example) that provides support for the user to use the signal strength collected by the terminal devices 1201 , 1202 , and 1203 . The background management server can analyze and process the received signal strength and other data, and feed back the processing results (such as sample sets, trained positioning models, and positioning results—just examples) to databases, automatic mobile devices, and so on.

It should be noted that the sample set construction method, the indoor positioning model construction method, and the indoor positioning method provided by the embodiments of the present invention are generally executed by the server 1205. Correspondingly, the sample set construction apparatus, the indoor positioning model construction apparatus, and the indoor positioning apparatus are generally Set in the server 1205.

It should be understood that the numbers of terminal devices, networks, servers, databases, and automated mobile devices in FIG. 12 are merely illustrative. There can be any number of terminal devices, networks, servers, databases, and automated mobile devices according to implementation needs.

Referring next to FIG. 13 , it shows a schematic structural diagram of a computer system 1300 suitable for implementing a terminal device or a server according to an embodiment of the present invention. The terminal device shown in FIG. 13 is only an example, and should not impose any limitations on the functions and scope of use of the embodiments of the present invention.

As shown in FIG. 13, a computer system 1300 includes a central processing unit (CPU) 1301, which can be loaded into a random access memory (RAM) 1303 according to a program stored in a read only memory (ROM) 1302 or a program from a storage section 1308 Instead, various appropriate actions and processes are performed. In the RAM 1303, various programs and data necessary for the operation of the system 1300 are also stored. The CPU 1301 , the ROM 1302 , and the RAM 1303 are connected to each other through a bus 1304 . An input/output (I/O) interface 1305 is also connected to bus 1304 .

The following components are connected to the I/O interface 1305: an input section 1306 including a keyboard, a mouse, etc.; an output section 1307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1308 including a hard disk, etc. ; and a communication section 1309 including a network interface card such as a LAN card, a modem, and the like. The communication section 1309 performs communication processing via a network such as the Internet. Drivers 1310 are also connected to I/O interface 1305 as needed. A removable medium 1311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is mounted on the drive 1310 as needed so that a computer program read therefrom is installed into the storage section 1308 as needed.

In particular, the processes described above with reference to the flowcharts may be implemented as computer software programs in accordance with the disclosed embodiments of the present invention. For example, embodiments disclosed herein include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the method illustrated in the flowchart. In such an embodiment, the computer program may be downloaded and installed from the network via the communication portion 1309, and/or installed from the removable medium 1311. When the computer program is executed by the central processing unit (CPU) 1301, the above-described functions defined in the system of the present invention are executed.

It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples of computer readable storage media may include, but are not limited to, electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, however, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code therein. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device . Program code embodied on a computer readable medium may be transmitted using any suitable medium including, but not limited to, wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more logical functions for implementing the specified functions executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams or flowchart illustrations, and combinations of blocks in the block diagrams or flowchart illustrations, can be implemented in special purpose hardware-based systems that perform the specified functions or operations, or can be implemented using A combination of dedicated hardware and computer instructions is implemented.

The units involved in the embodiments of the present invention may be implemented in a software manner, and may also be implemented in a hardware manner. The described unit may also be provided in the processor, for example, it may be described as: a processor includes a signal acquisition unit, a signal processing unit and a sample set construction unit. Wherein, the names of these units in some cases do not constitute a limitation of the unit itself, for example, the signal acquisition unit may also be described as "a unit for acquiring multiple signal strengths of known reference points".

As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or may exist alone without being assembled into the device. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by a device, the device includes: acquiring multiple signal strengths of known reference points; determining according to the multiple signal strengths High probability occurrence interval; select the signal strength belonging to the high probability occurrence interval from multiple signal intensities; use the signal strength belonging to the high probability occurrence interval to generate corresponding sample data for known reference points; use multiple known reference points Corresponding sample data, construct a sample set.

As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or may exist alone without being assembled into the device. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by a device, the device includes: acquiring a sample set, wherein the sample set is constructed from a plurality of sample data, and the sample data is composed of The signal strength belonging to the high probability occurrence interval is generated; using the sample set, iteratively trains the support vector machine to construct an indoor positioning model.

As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or may exist alone without being assembled into the device. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by a device, the device includes: acquiring the signal fingerprint information of the point to be located; inputting the signal fingerprint information into the indoor positioning model, Obtain the location information of the to-be-located point.

According to the technical solutions of the embodiments of the present invention, most of the acquired data, such as signal strength, are normal data, and the distribution of normal data is generally concentrated. A small part of abnormal data, that is, dirty data, is often outside the range or interval where normal data is located. Based on this, the solution provided by the embodiment of the present invention determines a high-probability occurrence interval according to multiple signal strengths; uses a high-probability occurrence interval The signal strength of the signal intensity is generated, and the corresponding sample data is generated for the known reference points; the sample data corresponding to multiple known reference points is used to construct the sample set, and the high probability occurrence interval is used to realize the removal of dirty data, that is, the sample data for constructing the sample set is composed of The data obtained after removing the dirty data can effectively remove the dirty data in the sample set.

The above-mentioned specific embodiments do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (17)

1. A sample set construction method is characterized by comprising the following steps:

acquiring a plurality of signal strengths of known reference points;

determining a high probability occurrence interval according to the signal intensities;

selecting the signal intensity belonging to the high probability occurrence interval from the plurality of signal intensities;

generating corresponding sample data for the known reference point by using the signal intensity belonging to the high-probability occurrence interval;

and constructing a sample set by using the sample data corresponding to the plurality of known reference points.

2. The sample set construction method of claim 1,

when the plurality of signal strengths originate from at least two signal sources, and each of the signal sources corresponds to a plurality of signal strengths,

and aiming at a plurality of signal strengths corresponding to each signal source, executing the step of determining the high-probability occurrence interval.

3. The method for constructing a sample set according to claim 2, wherein generating corresponding sample data for the known reference point comprises:

for a plurality of signal strengths belonging to the high probability occurrence interval and originating from the same signal source, performing: calculating an average value of the plurality of signal intensities, and taking the average value as sample data corresponding to the known reference point;

or,

for the signal strength belonging to the high probability occurrence interval, performing: and calculating the signal intensity difference between every two signal sources, and taking the signal intensity difference as sample data of the known reference point.

4. The sample set construction method according to claim 1 or 2, characterized by further comprising:

selecting a plurality of target known reference points which meet a preset distance threshold value for an unknown reference point from a plurality of known reference points;

assigning a corresponding weight to each of the target known reference points;

calculating corresponding sample data for the unknown reference point by using the weight and the sample data corresponding to the target known reference point;

adding sample data corresponding to the unknown reference point to the sample set.

5. A method for constructing an indoor positioning model is characterized by comprising the following steps:

acquiring a sample set, wherein the sample set is constructed by a plurality of sample data, and the sample data is generated by signal intensity belonging to a high probability occurrence interval;

and utilizing the sample set to iteratively train a support vector machine to construct an indoor positioning model.

6. The indoor positioning model construction method according to claim 5,

further comprising: setting a corresponding first value range for the parameters in the support vector machine;

randomly distributing initial parameters for the support vector machine in the first value range;

based on the initial parameters, performing the step of iteratively training a support vector machine.

7. The indoor positioning model construction method according to claim 6,

further comprising: setting an iteration step length for the parameters of the support vector machine;

the step of iteratively training the support vector machine comprises:

starting from the second iteration, taking each iteration as the current iteration, and executing:

determining a parameter corresponding to a previous iteration corresponding to the current iteration;

increasing or decreasing the iteration step length for the parameter corresponding to the previous iteration to serve as the parameter of the current iteration;

training the support vector machine based on the parameters of the current iteration;

and when the training result meets a preset first termination condition, terminating the iteration.

8. The indoor positioning model construction method according to claim 5 or 7,

further comprising: setting a corresponding second value range for the parameters in the support vector machine;

initializing a firework population in the second value range;

the step of iteratively training the support vector machine comprises:

for each iteration, performing:

determining a firework population, and determining explosion sparks and variant sparks of fireworks in the firework population;

calculating the fitness of the explosion spark and the variant spark;

judging whether the result of the current iteration meets a second termination condition,

if so, taking the explosion spark with the minimum fitness or the variation spark with the minimum fitness as a training result, and ending the current process;

otherwise, selecting a plurality of target sparks from the explosion sparks and the variation sparks to form a firework population, wherein the target sparks are used as fireworks in the firework population.

9. An indoor positioning method implemented based on an indoor positioning model constructed by any one of the methods of claims 5 to 8, comprising:

acquiring signal fingerprint information of a to-be-positioned point;

and inputting the signal fingerprint information into the indoor positioning model to obtain the position information of the to-be-positioned point.

10. The indoor positioning method of claim 9, wherein the acquiring the signal fingerprint information of the point to be positioned comprises:

the collected signal intensity of the point to be positioned is used as the signal fingerprint information;

or,

and calculating the signal intensity difference corresponding to each two signal sources according to the signal intensity of the acquired at least two signal sources at the point to be positioned, and taking the signal intensity difference as the signal fingerprint information.

11. A sample set constructing apparatus, comprising:

a signal acquisition unit for acquiring a plurality of signal strengths of known reference points;

the signal processing unit is used for determining a high-probability occurrence interval according to the signal intensities acquired by the signal acquisition unit; selecting a signal intensity belonging to the high probability occurrence section from the plurality of signal intensities;

the sample set constructing unit is used for generating corresponding sample data for the known reference point by utilizing the signal intensity which belongs to the high-probability occurrence interval and is selected by the signal processing unit; and constructing a sample set by using the sample data corresponding to the plurality of known reference points.

12. The sample set constructing apparatus according to claim 11,

the signal processing unit is configured to, when the plurality of signal strengths are derived from at least two signal sources and each of the signal sources corresponds to a plurality of signal strengths, perform, for the plurality of signal strengths corresponding to each of the signal sources, a step of determining a high-probability occurrence interval according to the plurality of signal strengths.

13. The sample set constructing apparatus according to claim 11 or 12,

the signal processing unit is further used for selecting a plurality of target known reference points which meet a preset distance threshold value for an unknown reference point from the known reference points; assigning a corresponding weight to each of the target known reference points;

the sample set constructing unit is further configured to calculate corresponding sample data for the unknown reference point by using the weight and the sample data corresponding to the target known reference point; adding sample data corresponding to the unknown reference point to the sample set.

14. An indoor positioning model building device, comprising:

the device comprises a sample set acquisition unit, a signal acquisition unit and a signal processing unit, wherein the sample set is constructed by a plurality of sample data, and the sample data is generated by signal intensity belonging to a high probability occurrence interval;

and the model construction unit is used for iteratively training a support vector machine by utilizing the sample set acquired by the sample set acquisition unit to construct an indoor positioning model.

15. An indoor positioning device, comprising:

the information acquisition unit is used for acquiring signal fingerprint information of a to-be-positioned point;

and the position determining unit is used for inputting the signal fingerprint information acquired by the information acquiring unit into the indoor positioning model to acquire the position information of the to-be-positioned point.

16. A sample set construction electronic device, comprising:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-4.

17. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-10.

Images