CN110794437B

CN110794437B - Satellite positioning signal strength prediction method and device

Info

Publication number: CN110794437B
Application number: CN201911049471.0A
Authority: CN
Inventors: 陈杰; 李坚强; 陈壮壮; 曾崛; 王云飞
Original assignee: Shenzhen Zhongke Baotai Technology Co ltd
Current assignee: Shenzhen Zhongke Baotai Aerospace Technology Co.,Ltd.
Priority date: 2019-10-31
Filing date: 2019-10-31
Publication date: 2021-10-22
Anticipated expiration: 2039-10-31
Also published as: CN110794437A

Abstract

The embodiment of the application discloses a method and a device for predicting satellite positioning signal strength, which are used for predicting the strength of a satellite positioning signal by acquiring the position information of a region to be predicted; and predicting a satellite positioning signal intensity predicted value corresponding to the position information by using the trained satellite positioning signal intensity prediction model, and then selecting a position point with the satellite positioning signal predicted value meeting a preset condition as an unmanned aerial vehicle flying point or landing point so as to ensure the suitability of the selected flying point or landing point. Like this, the problem that dependence manual work selected unmanned aerial vehicle departure point or landing point and consumed a large amount of time when having avoided unmanned aerial vehicle field work, simultaneously, still improved the security and the stability that unmanned aerial vehicle takes off or lands through selecting suitable departure point or landing point, ensured the security and the stability that unmanned aerial vehicle takes off or lands.

Description

Satellite positioning signal strength prediction method and device

Technical Field

The application belongs to the technical field of machine learning, and particularly relates to a satellite positioning signal strength prediction method and device.

Background

In the scene of autonomous flight of the unmanned aerial vehicle, selecting a proper flying start point or a proper landing point is a crucial condition.

At present, the unmanned aerial vehicle flying starting point is mainly selected by manually carrying the unmanned aerial vehicle, the satellite signal intensity of the position point where the unmanned aerial vehicle arrives is checked through a self-checking program of the unmanned aerial vehicle, and when the satellite signal intensity of the position point where the unmanned aerial vehicle arrives meets the condition, the position point is used as the flying starting point. And the selection of the unmanned aerial vehicle landing point is also to select a point suitable for landing mainly by means of artificial visual observation, for example, selecting an open space in an area as the landing point.

In addition, the unmanned aerial vehicle takeoff or landing point that is selected mainly by hand may not be the appropriate takeoff or landing point. If the selected flying starting point or landing point is not appropriate, the safety and stability of the takeoff or landing of the unmanned aerial vehicle can be influenced. For example, when the landing point is manually selected, the free space is directly used as the landing point, and the situation that the landing process is unstable may exist when the free space is used as the landing point, for example, the actual landing point has a large deviation from the selected landing point, so that the landing stability and safety of the unmanned aerial vehicle are affected.

In sum, the current selection mode of the takeoff point or the landing point of the unmanned aerial vehicle needs to consume a lot of time. In addition, the selected flying point or landing point is not necessarily a proper position point, so that the problems of instability and low safety still exist when the unmanned aerial vehicle takes off or lands.

Disclosure of Invention

The embodiment of the application provides a satellite positioning signal strength prediction method and device, and solves the problem that a large amount of time is wasted when an unmanned aerial vehicle depends on manual selection of a flying point or a landing point during field operation.

In a first aspect, an embodiment of the present application provides a method for predicting satellite positioning signal strength, including:

acquiring position information of an area to be predicted;

and predicting a satellite positioning signal strength predicted value corresponding to the position information by using the trained satellite positioning signal strength prediction model.

According to the method and the device, the satellite positioning signal strength corresponding to the position information is predicted by the satellite positioning signal strength prediction model through acquiring the position information, so that the flying point or the landing point of the unmanned aerial vehicle can be selected according to the predicted satellite positioning signal strength, a large amount of time is not consumed, the labor cost is reduced, and the selection efficiency of the flying point or the landing point of the unmanned aerial vehicle is improved.

With reference to the first aspect, in a possible implementation manner, after obtaining a satellite positioning signal strength predicted value corresponding to the location information, the method further includes:

and selecting a position point of which the satellite positioning signal intensity predicted value meets a preset condition as a flying point or a landing point of the unmanned aerial vehicle.

Further, the satellite positioning signal strength prediction value meets a preset condition, including:

the predicted value of the satellite positioning signal strength is greater than or equal to a preset threshold value;

and/or

And the predicted value of the satellite positioning signal strength is maximum.

With reference to the first aspect, in a possible implementation manner, the satellite positioning signal strength prediction model is a PSO-GPR model, and the PSO-GPR model is a PSO-pPITC model, a PSO-pPIC model, or a PSO-pICF model.

With reference to the first aspect, in a possible implementation manner, the predicting, by using a trained satellite positioning signal strength prediction model, a satellite positioning signal strength prediction value corresponding to the position information includes:

inputting the position information into the trained satellite positioning signal strength prediction model to obtain the variance output by the satellite positioning signal strength prediction model;

selecting a prediction point according to the variance;

and inputting the position information of the predicted point into the satellite positioning signal strength prediction model to obtain a satellite positioning signal strength predicted value of the predicted point output by the satellite positioning signal strength prediction model.

With reference to the first aspect, in a possible implementation manner, the training process of the satellite positioning signal strength prediction model includes:

acquiring training data, wherein the training data comprises position information of a target area and corresponding satellite positioning signal intensity data;

and training a pre-constructed satellite positioning signal strength prediction model by using the training data to obtain the trained satellite positioning signal strength prediction model and the optimal hyper-parameter.

With reference to the first aspect, in one possible implementation manner, the satellite positioning signal strength prediction model is a PSO-GPR model; the training data comprises training samples and test samples;

training a pre-constructed satellite positioning signal strength prediction model by using the training data to obtain a trained satellite positioning signal strength prediction model and an optimal hyper-parameter, wherein the training data comprises the following steps:

an initialization step: initializing particle swarm parameters of the PSO-GPR model;

and (3) intensity prediction step: based on the PSO-GPR model, calculating a satellite positioning signal intensity predicted value of the test sample according to the initialized particle swarm parameters, the test sample and the training sample;

and a fitness calculation step: calculating the fitness value of each particle according to the satellite positioning signal strength predicted value and the satellite positioning signal strength measured value of the test sample;

and (3) fitness evaluation step: comparing the fitness value of each particle with a historical optimal fitness value, determining a current generation optimal fitness value of each particle, and obtaining a current generation global optimal fitness value according to the current generation optimal fitness value of each particle and the optimal fitness value of a global particle;

a judging step: judging whether the maximum iteration times is reached; if not, updating the positions and the speeds of all the particles of the particle swarm, and returning to the strength prediction step, the fitness calculation step, the fitness evaluation step and the judgment step after generating a new particle swarm hyper-parameter; if so, obtaining an optimal hyper-parameter according to the current global optimal fitness value, and obtaining a trained satellite positioning signal strength prediction model.

With reference to the first aspect, in a possible implementation manner, the calculating a predicted value of a satellite positioning signal strength of the test sample according to the initialized particle swarm parameters, the test sample, and the training sample based on the PSO-GPR model includes:

dividing the training sample into a plurality of training subsamples, and distributing one training subsample to each node; wherein the number of training subsamples is equal to the number of nodes;

using a greedy algorithm to the test sample to obtain a support set, wherein the support set comprises position information;

based on the PSO-GPR model, according to the training subsample, the support set and the initialized particle swarm parameters, calculating a satellite positioning signal intensity predicted value of the test sample through the node and the main node; wherein the master node is a node randomly selected from the plurality of nodes.

With reference to the first aspect, in one possible implementation manner, the PSO-GPR model is a PSO-pPITC model;

based on the PSO-GPR model, according to the training subsample, the support set and the initialized particle swarm parameters, the satellite positioning signal intensity predicted value of the test sample is calculated through the node and the main node, and the method comprises the following steps:

controlling each node to calculate a local abstract tuple of each node according to the support set, the training subsample and the hyper-parameter, and sending the local abstract tuple to the main node; wherein the initialized particle swarm parameters comprise the hyper-parameters;

controlling the master node to calculate a global digest tuple according to the local digest tuple;

and controlling the main node to calculate the satellite positioning signal strength predicted value of each test sample according to the support set, the global abstract tuple and the test samples.

With reference to the first aspect, in one possible implementation manner, the PSO-GPR model is a PSO-pPIC model;

controlling each node to calculate a local abstract tuple according to the support set, the training subsample and the hyper-parameter; wherein the initialized particle swarm parameters comprise the hyper-parameters;

controlling each node to send the local digest tuple to the master node to instruct the master node to calculate a global digest tuple according to the local digest tuple and send the global digest tuple to each node;

calculating a cluster center point of each node by using a clustering algorithm on the training subsample of each node;

and dividing the test sample to corresponding target nodes according to the clustering center point to indicate each target node to calculate the satellite positioning signal strength predicted value of the test sample according to the test sample, the local abstract tuple and the global abstract tuple obtained by division.

With reference to the first aspect, in a possible implementation manner, the dividing the test samples into corresponding target nodes according to the clustering center point includes:

respectively calculating the distance between each sample in the test samples and each cluster central point;

and for each sample, taking the node corresponding to the minimum distance as a target node corresponding to the sample, and dividing the sample into corresponding target nodes.

With reference to the first aspect, in one possible implementation manner, the PSO-GPR model is a PSO-pICF model;

controlling each node to calculate the factor of the covariance matrix of each node through an ICF theory according to the support set and the hyperparameter; wherein the initialized particle swarm parameters comprise the hyper-parameters;

controlling each node to respectively calculate a local abstract tuple of each node according to the training subsample and the factor of the covariance matrix, and sending the local abstract tuple to the main node;

controlling the main node to calculate a global summary tuple according to the local summary tuple and sending the global summary tuple to each node;

controlling each node to calculate a prediction component tuple according to the training subsample, the global summary tuple and a target component of the local summary tuple, and sending the prediction component tuple to the master node;

and controlling the main node to calculate the satellite positioning signal strength predicted value of the test sample according to the prediction component tuple.

In a second aspect, an embodiment of the present application provides a satellite positioning signal strength prediction apparatus, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method of any one of the above first aspects when executing the computer program.

In a third aspect, an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored, and the computer program, when executed by a processor, implements the method of any one of the above first aspects.

In a fourth aspect, embodiments of the present application provide a computer program product, which, when run on a terminal device, causes a satellite positioning signal strength prediction apparatus to perform the method of any one of the first aspect.

It is understood that the beneficial effects of the second to fourth aspects can be seen from the description of the first aspect, and are not described herein again.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

Fig. 1 is a schematic block diagram of a training process of a satellite positioning signal strength prediction model according to an embodiment of the present disclosure;

fig. 2 is a schematic block diagram of a specific flow of step S102 provided in an embodiment of the present application;

fig. 3 is a schematic block diagram of a flow of step S202 provided in the embodiment of the present application;

fig. 4 is a schematic block diagram of a flow of step S303 provided in the embodiment of the present application;

fig. 5 is another schematic block diagram of the flow of step S303 provided in the embodiment of the present application;

fig. 6 is a schematic block diagram of another flowchart of step S303 provided in the embodiment of the present application;

fig. 7 is a schematic block diagram of a flow chart of a method for predicting satellite positioning signal strength according to an embodiment of the present application;

fig. 8 is a schematic block diagram of a flow of a method for selecting a takeoff point or a landing point of an unmanned aerial vehicle according to an embodiment of the present application;

fig. 9 is a block diagram of a satellite positioning signal strength predicting apparatus according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of an apparatus for predicting satellite positioning signal strength according to an embodiment of the present disclosure.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application.

According to the embodiment of the application, the satellite positioning signal intensity corresponding to the position information is predicted by acquiring the position information of the area to be predicted and then using a pre-trained satellite positioning signal intensity prediction model. After the satellite positioning signal intensity of the corresponding position is predicted, the unmanned aerial vehicle flying point or landing point can be selected according to the satellite positioning signal intensity. Compared with the existing mode of selecting the flying point or the landing point through the mode of manually traversing each position in the area, the scheme provided by the embodiment of the application can reduce the labor cost and shorten the time. In addition, according to the satellite positioning signal intensity predicted values of the position points, the position points with the predicted values meeting the preset conditions are selected to serve as the takeoff point or the landing point of the unmanned aerial vehicle, so that the suitability of the selected takeoff point or the landing point is ensured, and the takeoff or landing safety and the landing stability of the unmanned aerial vehicle are ensured.

The scheme provided by the embodiment of the application is described in a sub-model training phase and a prediction phase.

Model training phase

Referring to fig. 1, a schematic block diagram of a training process of a satellite positioning signal strength prediction model provided in an embodiment of the present application, where the training process may include the following steps:

step S101, training data are obtained, wherein the training data comprise position information of a target area and corresponding satellite positioning signal intensity data.

The target region may be any region. That is, the target region may be a region to be predicted or a region other than the region to be predicted. For example, the region to be predicted may be an a region, and the target region may be an a region or a B region different from the a region. Compared with the method, the target area is the same as the area needing to be predicted, and the prediction result can be more accurate.

The position information may be coordinates and/or longitude and latitude information. Typically, there are a plurality of points within the target area and the location information includes location information for corresponding points within the target area. For example, the training data includes latitude and longitude information of all points in the a area.

The satellite positioning signal strength data may be, for example, but not limited to, GPS signal strength data, GLONASS signal strength data, Galileo signal strength data, or beidou satellite positioning signal strength data.

It will be appreciated that each point has its corresponding satellite positioning signal strength. For example, the training data includes points a, B, C, D, and E in the target area, and the GPS signal strengths corresponding to these 5 points.

The training data can be acquired through an acquisition module, and the acquisition module acquires the position information and the satellite positioning signal intensity corresponding to the position information. The acquisition module can be a module consisting of a GPS unit and a serial port, and can also be a module for installing GPStest software. In a specific application, the acquisition module may be specifically installed on an unmanned vehicle or other mobile devices, for example, when the acquisition module is installed on the unmanned vehicle, the unmanned vehicle is controlled to traverse the target area, so as to acquire the position information of the target area and the corresponding satellite positioning signal strength. For another example, when the acquisition module is installed on a mobile phone, the mobile phone can be moved in the target area to acquire the required training data.

Step S102, training a pre-constructed satellite positioning signal strength prediction model by using training data to obtain the trained satellite positioning signal strength prediction model and the optimal hyper-parameter.

The satellite positioning signal strength prediction model is used for predicting satellite positioning signal strength, and the input of the model is position information and the output is a satellite positioning signal strength value corresponding to the position information. For example, the latitude and longitude of the point A are input into the model, and the GPS signal intensity value of the point A is output from the model.

The satellite positioning signal strength prediction model can be, but is not limited to, a Particle Swarm Optimization-based Gaussian process regression model PSO-GPR model (GPR, Gaussian process regression, PSO).

Further, referring to the specific flow schematic block diagram of step S102 shown in fig. 2, the satellite positioning signal strength prediction model is a PSO-GPR model; the training data includes training samples and test samples, wherein the training samples and the test samples both include position information and corresponding satellite positioning signal strengths.

The step S102, namely, the process of training the pre-constructed satellite positioning signal strength prediction model by using the training data to obtain the trained satellite positioning signal strength prediction model and the optimal hyper-parameter may include:

step S201, an initialization step, which specifically includes: particle swarm parameters of the PSO-GPR model are initialized.

The particle swarm parameters include a particle swarm size, an inertial weight, a search area (i.e., a maximum position) of the parameters, a maximum speed, a maximum number of iterations, a hyper-parameter, and the like.

The above-mentioned hyper-parameters refer to parameters of the kernel function. There are various kernel functions, and before initializing particle swarm parameters, a Gaussian kernel function can be artificially selected as the kernel function according to actual needs. The calculation formula of the gaussian kernel function is as follows:

wherein x is₁Is any point in space, x₂Is the kernel center and σ is a parameter of the kernel.

Of course, the kernel function may be other functions, and is not limited to the gaussian kernel function mentioned above.

Step S202, an intensity prediction step, which specifically comprises the following steps: and calculating the satellite positioning signal intensity predicted value of the test sample according to the initialized particle swarm parameters, the test sample and the training sample based on the PSO-GPR model.

Specifically, after initializing the particle swarm parameters, a PSO-GPR model may be trained according to the initialized particle swarm parameters and the training sample, and after the training is completed, the PSO-GPR model may be used to calculate a predicted value of the satellite positioning signal strength of the test sample.

In some embodiments, referring to the schematic block diagram of the flow of step S202 shown in fig. 3, the specific flow of step S202 may include:

step S301, dividing a training sample into a plurality of training subsamples, and distributing one training subsample for each node; wherein the number of training subsamples is equal to the number of nodes.

It should be noted that, the training samples may be divided into m parts on average, and then the m training subsamples are divided into m nodes respectively; or randomly dividing the training sample into m parts, and then dividing the m parts into m nodes. Wherein one node corresponds to one piece of training subsample data.

Step S302, a greedy algorithm is used for the test sample to obtain a support set, and the support set comprises position information.

The position information of the support set is the position information of the test sample. Greedy algorithms are well known to those skilled in the art and will not be described in detail herein.

Step S303, based on a PSO-GPR model, calculating a satellite positioning signal intensity predicted value of a test sample through a node and a main node according to parameters of a training subsample, a support set and initialized particle swarm; the master node is a node randomly selected from a plurality of nodes.

The PSO-GPR model is a PSO-pPITC model, a PSO-pPIC model or a PSO-pICF model. Specifically, based on a support set of each node, a training subsample, initialized particle swarm parameters and the like, the satellite positioning signal strength predicted value of the test sample is calculated in parallel through a plurality of nodes and a main node.

The satellite positioning signal strength prediction process of the test sample of the PSO-pPITC model, the PSO-pPIC model and the PSO-pICF model will be described below.

The PSO-GPR model is a PSO-pPITC model

Referring to a schematic block diagram of a flow of step S303 shown in fig. 4, a specific process of step S303 may include:

s401, controlling each node to calculate a local abstract tuple of each node according to the support set, the training subsamples and the hyper-parameters, and sending the local abstract tuple to the main node; the initialized particle swarm parameters comprise hyper-parameters.

Specifically, one node is randomly selected from a plurality of nodes as a master node. And then each node is controlled to calculate the local abstract, namely each node calculates the local abstract tuple of itself according to the support set, the training subsample and the hyper-parameter of itself and the like. After each node calculates the local abstract of the node, each node sends the local abstract to the main node.

The calculation formula of the local summary may be as follows (2) and (3).

Wherein the training sample is (D, y)_D) Training subsamples of each node are

S represents a support set, and the local abstract tuple is

And S402, the control main node calculates a global summary tuple according to the local summary tuple.

Specifically, after each node sends its local digest tuple to the master node, the master node calculates a global digest tuple from the received local digest tuple and the following equations (4) and (5)

And S403, the control main node calculates the satellite positioning signal strength predicted value of each test sample according to the support set, the global abstract tuple and the test samples.

Specifically, after calculating the global summary tuple, the master node may calculate a satellite positioning signal strength predicted value and a variance corresponding to each location information of the test sample according to the support set S, the global summary tuple, and the location information of the test sample, where the calculation formulas are as follows (6) and (7)

Wherein u is_mIn order to test the sample, the sample is,

the satellite positioning signal strength predictions corresponding to the respective location information represented as the test samples,

the prediction variance corresponding to each piece of location information is indicated.

The PSO-GPR model is a PSO-pPIC model

When the satellite positioning signal strength of the test sample is predicted through the PSO-pPITC model, the prediction effect may be poor, and based on the situation, prediction can be performed based on the PSO-pPIC model in order to improve the prediction effect.

Referring to another schematic flow chart of step S303 shown in fig. 5, the specific process of step S303 may include:

s501, controlling each node to calculate a local abstract tuple according to the support set, the training subsample and the hyper-parameter; the initialized particle swarm parameters comprise hyper-parameters.

Specifically, a node is randomly selected from a plurality of nodes to serve as a master node, and then each node is controlled to calculate a local summary tuple. Each node calculates a local summary from the support set, training subsamples and hyper-parameters by equations (2) and (3) above.

And step S502, controlling each node to send the local summary tuples to the master node so as to instruct the master node to calculate the global summary tuples according to the local summary tuples and send the global summary tuples to each node.

Specifically, after each node calculates its own local digest tuple, it sends the local digest tuple to the master node. The master node calculates a global digest tuple according to the received local digest tuple through the above formulas (4) and (5), and transmits the calculated global digest tuple to each node.

And S503, calculating the clustering center point of each node by using a clustering algorithm for the training subsample of each node.

Specifically, for a node, the clustering center point of the node is calculated through a clustering algorithm according to the position information of the training subsample of the node. For example, according to the longitude and latitude information of the training subsample divided by a node, a clustering center point of the node is obtained by using a clustering algorithm. And calculating the clustering center point of each node respectively according to the calculation result.

And step S504, dividing the test sample to corresponding target nodes according to the clustering center point to indicate each target node to calculate the satellite positioning signal strength predicted value of the test sample according to the test sample, the local abstract tuple and the global abstract tuple obtained through division.

Specifically, after the clustering center point of each node is determined, the test sample may be divided into corresponding nodes according to each clustering center point. The test sample comprises a plurality of pieces of position information, the distance from one piece of position information to each clustering center point is calculated, and then the clustering center point with the minimum distance value is used as a target node corresponding to the position information, namely, the sample corresponding to the position information is divided into the target nodes.

For example, the test sample comprises longitude and latitude information of a point A, a point B, a point C, a point E, a point D, a point E and corresponding GPS signal intensity values, for the point A, the Euclidean distance from the point A to each clustering center point is calculated according to the longitude and latitude information of the point A and the longitude and latitude information of each clustering center point, then the calculated Euclidean distances are sequenced, the clustering center point with the minimum Euclidean distance value is used as a target node corresponding to the point A, and sample data corresponding to the point A is divided into the corresponding target nodes. Similarly, the Euclidean distances from the point B, the point C, the point D and the point E to each clustering center point are respectively calculated, then the node with the minimum Euclidean distance is used as a target node, and the sample data of the point is divided into corresponding target nodes so as to divide all the sample data of the test sample into corresponding nodes.

That is, the specific process of dividing the test sample into the corresponding target nodes according to the cluster center point may include: respectively calculating the distance between each sample in the test samples and each cluster central point; and for each sample, taking the node corresponding to the minimum distance as a target node corresponding to the sample, and dividing the sample into the corresponding target nodes.

It can be understood that, each sample data is divided into corresponding nodes according to the distance between the test sample and the cluster center point, and there may be a case that some nodes do not divide sample data, some nodes divide many sample data, or some nodes divide few sample data.

After the test sample is divided into the corresponding target nodes, each target node may calculate a satellite positioning signal strength prediction value corresponding to the test sample by using the divided test sample data according to the global summary tuple sent by the master node and the local summary tuple corresponding to the target node, where the specific calculation formulas are as follows (8), (9) and (10).

The PSO-GPR model is a PSO-pICF model;

referring to still another schematic flow chart of step S303 shown in fig. 6, the specific process of step S303 may include:

s601, controlling each node to calculate the factor of the covariance matrix of each node through an ICF (incomplete matrix decomposition) theory according to the support set and the hyperparameter; the initialized particle swarm parameters comprise hyper-parameters.

Specifically, one node is randomly selected from a plurality of nodes as a master node. Each node is then controlled to calculate the factors of the covariance matrix of each node according to the support set and the hyperparameters using ICF (incomplete matrix decomposition) theory and equation (11).

And step S602, controlling each node to respectively calculate the local abstract tuple of each node according to the training subsample and the factor of the covariance matrix, and sending the local abstract tuple to the main node.

Specifically, after each node calculates the factor of the covariance matrix, each node is controlled to train the subsample according to the node

And the factor F of the covariance matrix_mCalculating the local digest tuple by equations (12), (13) and (14)

And step S603, the control main node calculates a global summary tuple according to the local summary tuple and sends the global summary tuple to each node.

Specifically, after each node is controlled to calculate a local digest tuple, each node sends the respective local digest tuple to the master node, and the master node calculates a global digest tuple according to the received local digest tuple through formulas (15) and (16). And then the master node sends the global summary tuple to each node.

Wherein the global summary tuple is

Step S604, controlling each node to be according to target components of the training subsamples, the global summary tuples and the local summary tuples

Computing prediction component tuples

And sends the prediction component tuple to the master node.

Specifically, after receiving the global digest tuple, each node calculates a respective prediction component tuple according to one target component of the training subsample, the global digest tuple, and the local digest tuple through equations (17) and (18). Each node then sends the respective prediction component tuple to the master node.

And step S605, the control main node calculates the satellite positioning signal intensity predicted value of the test sample according to the prediction component tuple.

Specifically, after receiving the prediction component tuples of each node, the master node calculates the predicted values of the satellite positioning signal strength corresponding to each position information of the test sample through formulas (19) and (20)

Sum variance

It can be seen that when the satellite positioning signal strength predicted value of the test sample is calculated, a plurality of nodes are set for parallel calculation, the calculation time is shortened, and the calculation efficiency is improved.

After the predicted value of the satellite positioning signal strength of the test sample is calculated, the fitness of the particles can be calculated according to the predicted value and the measured value.

Step S203, a fitness calculation step, which specifically includes: and calculating the fitness value of each particle according to the satellite positioning signal strength predicted value and the satellite positioning signal strength measured value of the test sample.

Specifically, the fitness value of each particle is calculated according to formula (21).

Where f is the fitness value of each particle, { g (x)_j) J is the predicted satellite positioning signal strength value of all test samples, { y ═ 1,2_jJ 1,2,3, k is the actual satellite positioning signal strength value of all test samples.

Step S204, a fitness evaluation step, which specifically comprises the following steps: and comparing the fitness value of each particle with the historical optimal fitness value, determining the current generation optimal fitness value of each particle, and obtaining the current generation global optimal fitness value according to the current generation optimal fitness value of each particle and the optimal fitness value of the global particle.

The historical optimal fitness value is the optimal fitness value in the previous iteration process. The calculation formula of the current generation global optimum fitness value is formula (22).

g_hesl ^m+1＝arg min(f(g_best ^m)，f(P_hesti ^m+1)) (22)

Step S205, a judging step, which specifically comprises: judging whether the maximum iteration times is reached; if not, updating the positions and the speeds of all the particles of the particle swarm, and returning to the strength prediction step, the fitness calculation step, the fitness evaluation step and the judgment step after generating a new particle swarm hyper-parameter; if so, obtaining the optimal hyper-parameter according to the current global optimal fitness value, and obtaining a trained satellite positioning signal strength prediction model.

Specifically, if the maximum iteration number is reached, the hyper-parameter and the trained model at the moment are obtained, and the hyper-parameter at the moment is the optimal hyper-parameter. Otherwise, if the maximum iteration number is not reached, after updating the corresponding parameters, the steps S202, S203, S204 and S205 are returned again until the maximum iteration number is reached.

Prediction phase

After the model training is completed, the trained satellite positioning signal strength prediction model can be used for prediction. The prediction process will be described below.

Referring to fig. 7, a schematic flow chart of a method for predicting satellite positioning signal strength according to an embodiment of the present application may include the following steps:

step S701, position information of the area to be predicted is obtained.

It should be noted that the location information may be latitude and longitude information and/or coordinate information. The position information may specifically pass through an acquisition module, which may specifically be installed on an unmanned vehicle or other mobile device.

The area to be predicted may include a plurality of position points, and the obtaining of the satellite information of the area to be predicted may be obtaining of position information of all position points in the area to be predicted, or obtaining of position information of a part of position points in the area to be predicted.

And S702, predicting a satellite positioning signal strength predicted value corresponding to the position information by using the trained satellite positioning signal strength prediction model.

Specifically, the acquired position information is input to a trained satellite positioning signal strength prediction model, and a satellite positioning signal strength prediction value output by the satellite positioning signal strength prediction model is obtained. For example, the latitude and longitude information of a certain point is input into a satellite positioning signal strength prediction model, and a predicted value of the satellite positioning signal strength of the point is output.

The satellite positioning signal strength prediction model can be, but is not limited to, a PSO-GPR model, which is a PSO-pPITC model, a PSO-pPIC model or a PSO-pICF model.

It is understood that after the training of the model is completed, parameters of the model, such as local digest tuples, global digest tuples, support sets, hyper-parameters, and kernel functions, are determined. In the process of prediction by using the model, the calculation process is similar to the corresponding calculation process of the training process, except that the local summary tuples, the global summary tuples, the support sets and the hyper-parameters and the like are not required to be recalculated.

For example, when the PSO-GPR model is a PSO-pPITC model, a satellite positioning signal strength prediction value corresponding to the position information is calculated according to the support set, the global digest tuple, and the input position information.

Therefore, the satellite positioning signal strength corresponding to the position information is predicted by the satellite positioning signal strength prediction model through acquiring the position information, so that the flying point or the landing point of the unmanned aerial vehicle can be selected according to the predicted satellite positioning signal strength, a large amount of time is not consumed, and the labor cost is reduced.

After the satellite positioning signal strength values of the corresponding areas are predicted, the corresponding departure points can be selected according to the preset departure conditions or the landing points can be selected according to the landing conditions.

Referring to fig. 8, a schematic flow chart of a method for selecting a takeoff point or a landing point of an unmanned aerial vehicle may include:

step S801, position information of the area to be predicted is acquired.

And S802, predicting a satellite positioning signal strength predicted value corresponding to the position information by using the trained satellite positioning signal strength prediction model.

It should be noted that steps S801 to S802 are the same as steps S701 to S702, and the related description refers to the above corresponding contents, which are not repeated herein.

And S803, selecting a position point of which the satellite positioning signal strength predicted value meets a preset condition as a flying point or a landing point of the unmanned aerial vehicle.

It should be noted that the preset conditions are set to ensure the suitability of the selected takeoff point or landing point, so as to ensure the safety and stability of takeoff or landing of the unmanned aerial vehicle. In some embodiments, the preset condition may be that the predicted value of the satellite positioning signal strength is the largest, that is, a position point with the largest predicted value of the satellite positioning signal strength is selected as a departure point or a landing point of the drone.

However, in some cases, the surrounding environment of the location point with the largest predicted value of the satellite positioning signal strength may not be suitable for the unmanned aerial vehicle to take off or land, for example, the location point with the largest predicted value of the satellite positioning signal strength is a steep slope or a lake, and at this time, the location point cannot be selected as a landing point or a takeoff point.

In some other embodiments, the predetermined condition may be that the predicted value of the satellite positioning signal strength is greater than or equal to a predetermined threshold, that is, a point where the predicted value of the satellite positioning signal strength is greater than or equal to the prediction threshold is taken as a candidate position point, and then a point with a greater strength or a point with a maximum strength is selected from the candidate position points as a start point or a landing point as needed.

It should be noted that the preset threshold may be set according to actual needs, for example, when an unmanned aerial vehicle departure point needs to be selected, the preset threshold is 28db, that is, when a predicted value of satellite positioning signal strength of a certain point is greater than or equal to 28db, the point is selected as the unmanned aerial vehicle departure point or a candidate departure point, and otherwise, position information of other areas is obtained for prediction.

Therefore, based on the satellite positioning signal intensity predicted value of each position point, a target position point is selected from the position points according to preset conditions, and the target position point is used as a flying point or a landing point of the unmanned aerial vehicle, so that the suitability of the selected flying point or landing point is ensured, and the safety and the stability of the unmanned aerial vehicle in taking off or landing are ensured.

In the prediction process, the collected position information can be all input into the satellite positioning signal strength prediction model to predict the satellite positioning signal strength prediction values corresponding to all the position information. Or the variance of all the position information can be calculated first, and then the position information needing to predict the satellite positioning signal intensity can be selected according to the variance.

In some embodiments, the specific process of predicting the satellite positioning signal strength predicted value corresponding to the position information by using the trained satellite positioning signal strength prediction model may include:

inputting the position information into a trained satellite positioning signal strength prediction model to obtain a variance output by the satellite positioning signal strength prediction model; selecting a prediction point according to the variance; and inputting the position information of the predicted point into a satellite positioning signal strength prediction model to obtain a satellite positioning signal strength predicted value of the predicted point output by the satellite positioning signal strength prediction model.

Specifically, the trained satellite positioning signal strength prediction model is used for calculating the variance corresponding to each position information, the variance represents the uncertainty of the position point, and the larger the variance is, the larger the uncertainty is. After calculating the variance corresponding to each piece of position information, selecting a point with a large variance value as a point with a large uncertainty factor, wherein the point with the large uncertainty factor is a point needing to be predicted, predicting the satellite positioning signal strength value of the selected predicted point, comparing the satellite positioning signal strength values of a plurality of predicted points, and finally determining the starting point or the landing point of the unmanned aerial vehicle.

After the variance is calculated, when a prediction point is selected according to the variance, the prediction point can be selected according to the size of a set threshold value, namely, a position point with the variance larger than the set threshold value is used as the prediction point, and the threshold value is set according to needs; and a plurality of position points with larger variance can be selected as the prediction points according to the requirement.

It can be seen that in this way, intensity prediction is not performed on each point of the whole region to be predicted, and compared with intensity prediction performed on all points, the method reduces the amount of calculation, thereby improving the prediction efficiency.

It should be noted that, the acquired training data is, for example, data of the a region, the data of the a region is used to train the model, and the finally obtained satellite positioning signal strength prediction model may be used for the satellite positioning signal strength of the a region, or may be directly used for predicting the satellite positioning signal strength of other regions (regions outside the a region), but the prediction may be inaccurate. Therefore, when the satellite positioning signal strength of other regions except the region a is to be predicted, it is preferable to collect data of the region to be predicted to retrain the satellite positioning signal strength prediction model.

Fig. 8 shows a block diagram of a satellite positioning signal strength predicting apparatus according to an embodiment of the present application, which corresponds to the satellite positioning signal strength predicting method according to the foregoing embodiment.

Referring to fig. 9, the apparatus includes:

a position information obtaining module 91, configured to obtain position information of an area to be predicted;

and the prediction module 92 is configured to calculate a satellite positioning signal strength prediction value corresponding to the position information by using the trained satellite positioning signal strength prediction model.

In a possible implementation manner, the apparatus may further include:

and the selecting module 93 is used for selecting a position point of which the satellite positioning signal strength predicted value meets a preset condition as a flying point or a landing point of the unmanned aerial vehicle. Further, the satellite positioning signal strength predicted value meets the preset conditions, including: the predicted value of the satellite positioning signal intensity is greater than or equal to a preset threshold value; and/or the satellite positioning signal strength prediction value is maximized.

In one possible implementation, the satellite positioning signal strength prediction model is a PSO-GPR model, and the PSO-GPR model is a PSO-pPITC model, a PSO-pPIC model or a PSO-pICF model.

In a possible implementation manner, the prediction module is specifically configured to:

inputting the position information into a trained satellite positioning signal strength prediction model to obtain a variance output by the satellite positioning signal strength prediction model;

selecting a prediction point according to the variance;

and inputting the position information of the predicted point into a satellite positioning signal strength prediction model to obtain a satellite positioning signal strength predicted value of the predicted point output by the satellite positioning signal strength prediction model.

In a possible implementation manner, the apparatus further includes a training module 94, where the training module is configured to:

training a pre-constructed satellite positioning signal strength prediction model by using training data to obtain the trained satellite positioning signal strength prediction model and the optimal hyper-parameter.

Furthermore, the satellite positioning signal intensity prediction model is a PSO-GPR model; the training data comprises training samples and test samples;

the model training module is specifically configured to perform:

an initialization step: initializing particle swarm parameters of a PSO-GPR model;

and (3) intensity prediction step: based on a PSO-GPR model, calculating a satellite positioning signal intensity predicted value of the test sample according to the initialized particle swarm parameters, the test sample and the training sample;

and (3) fitness evaluation step: comparing the fitness value of each particle with the historical optimal fitness value, determining the current generation optimal fitness value of each particle, and obtaining the current generation global optimal fitness value according to the current generation optimal fitness value of each particle and the optimal fitness value of the global particle;

a judging step: judging whether the maximum iteration times is reached; if not, updating the positions and the speeds of all the particles of the particle swarm, and returning to the strength prediction step, the fitness calculation step, the fitness evaluation step and the judgment step after generating a new particle swarm hyper-parameter; if so, obtaining the optimal hyper-parameter according to the current global optimal fitness value, and obtaining a trained satellite positioning signal strength prediction model.

Further, in a possible implementation manner, the model training module is specifically configured to:

a greedy algorithm is used for the test sample to obtain a support set, and the support set comprises position information;

based on a PSO-GPR model, calculating a satellite positioning signal intensity predicted value of a test sample through nodes and a main node according to parameters of a training sub-sample, a support set and initialized particle swarm; the master node is a node randomly selected from a plurality of nodes.

Furthermore, when the PSO-GPR model is a PSO-pPITC model, the training module is specifically configured to:

controlling each node to calculate a local abstract tuple of each node according to the support set, the training subsample and the hyper-parameter, and sending the local abstract tuple to the main node; the initialized particle swarm parameters comprise hyper-parameters;

the control main node calculates a global abstract tuple according to the local abstract tuple;

and the control main node calculates the satellite positioning signal intensity predicted value of each test sample according to the support set, the global abstract tuple and the test samples.

Furthermore, when the PSO-GPR model is a PSO-pPIC model, the training module is specifically configured to:

controlling each node to calculate a local abstract tuple according to the support set, the training subsample and the hyper-parameter; the initialized particle swarm parameters comprise hyper-parameters;

controlling each node to send the local abstract tuple to the main node so as to instruct the main node to calculate a global abstract tuple according to the local abstract tuple and send the global abstract tuple to each node;

calculating the clustering center point of each node by using a clustering algorithm on the training subsample of each node;

In a possible implementation manner, the training module is specifically configured to:

and for each sample, taking the node corresponding to the minimum distance as a target node corresponding to the sample, and dividing the sample into the corresponding target nodes.

Further, when the PSO-GPR model is a PSO-pICF model, the training module is specifically configured to:

controlling each node to calculate the factor of the covariance matrix of each node through an ICF theory according to the support set and the hyperparameter; the initialized particle swarm parameters comprise hyper-parameters;

the control main node calculates a global abstract tuple according to the local abstract tuple and sends the global abstract tuple to each node;

controlling each node to calculate a prediction component tuple according to the target components of the training subsample, the global abstract tuple and the local abstract tuple, and sending the prediction component tuple to the main node;

and the control main node calculates the satellite positioning signal intensity predicted value of the test sample according to the prediction component tuple.

It should be noted that, for the information interaction, execution process, and other contents between the above-mentioned devices/units, the specific functions and technical effects thereof are based on the same concept as those of the embodiment of the method of the present application, and specific reference may be made to the part of the embodiment of the method, which is not described herein again.

Fig. 10 is a schematic structural diagram of an apparatus for predicting satellite positioning signal strength according to an embodiment of the present disclosure. As shown in fig. 10, the satellite positioning signal strength prediction apparatus 10 of this embodiment includes: at least one processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the at least one processor 100, the processor 100 implementing the steps in any of the various satellite positioning signal strength prediction method embodiments described above when executing the computer program 102.

The satellite positioning signal strength predicting apparatus 10 may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server, and the satellite positioning signal strength predicting apparatus may be integrated with a device such as an unmanned vehicle. Of course, the satellite positioning signal strength prediction device may also be embodied as an unmanned aerial vehicle or other mobile equipment. The satellite positioning signal strength predicting device may include, but is not limited to, a processor 100 and a memory 101. It will be understood by those skilled in the art that fig. 10 is only an example of the satellite positioning signal strength predicting apparatus 10, and does not constitute a limitation to the satellite positioning signal strength predicting apparatus 10, and may include more or less components than those shown, or combine some components, or different components, for example, may also include an input-output device, a network access device, and the like.

The Processor 100 may be a Central Processing Unit (CPU), and the Processor 100 may be other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, a discrete hardware component, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 101 may be an internal storage unit of the satellite positioning signal strength predicting apparatus 10 in some embodiments, for example, a hard disk or a memory of the satellite positioning signal strength predicting apparatus 10. The memory 101 may also be an external storage device of the satellite positioning signal strength predicting apparatus 10 in other embodiments, for example, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), or the like provided on the satellite positioning signal strength predicting apparatus 10. Further, the memory 101 may also include both an internal storage unit and an external storage device of the satellite positioning signal strength prediction apparatus 10. The memory 101 is used for storing an operating system, an application program, a BootLoader (BootLoader), data, and other programs, such as program codes of the computer program. The memory 101 may also be used to temporarily store data that has been output or is to be output.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

The embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the computer program implements the steps in the above-mentioned method embodiments.

Embodiments of the present application provide a computer program product, which when running on a satellite positioning signal strength prediction apparatus, enables the satellite positioning signal strength prediction apparatus to implement the steps in the above-mentioned method embodiments when executed.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium and can implement the steps of the embodiments of the methods described above when the computer program is executed by a processor. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer readable medium may include at least: any entity or device capable of carrying computer program code to a photographing apparatus/terminal apparatus, a recording medium, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), an electrical carrier signal, a telecommunications signal, and a software distribution medium. Such as a usb-disk, a removable hard disk, a magnetic or optical disk, etc. In certain jurisdictions, computer-readable media may not be an electrical carrier signal or a telecommunications signal in accordance with legislative and patent practice.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.

Claims

1. A method for predicting satellite positioning signal strength, comprising:

acquiring position information of an area to be predicted;

predicting a satellite positioning signal intensity predicted value corresponding to the position information by using a trained satellite positioning signal intensity prediction model;

the predicting the satellite positioning signal strength predicted value corresponding to the position information by using the trained satellite positioning signal strength prediction model comprises the following steps:

inputting the position information of the area to be predicted into the trained satellite positioning signal strength prediction model to obtain the variance output by the satellite positioning signal strength prediction model;

selecting a prediction point according to the variance;

2. The method of claim 1, further comprising, after obtaining the predicted value of the satellite positioning signal strength corresponding to the position information:

3. The method according to claim 2, wherein the step of predicting the satellite positioning signal strength comprises:

and/or

4. The method according to any one of claims 1 to 3, wherein the training process of the satellite positioning signal strength prediction model comprises:

5. The method of claim 4, wherein the satellite positioning signal strength prediction model is a PSO-GPR model; the training data comprises training samples and test samples;

6. The method according to claim 5, wherein the calculating a predicted value of the satellite positioning signal strength of the test sample according to the initialized particle swarm parameters, the test sample and the training sample based on the PSO-GPR model comprises:

7. The method of satellite positioning signal strength prediction according to claim 6,

8. The method of satellite positioning signal strength prediction according to claim 6,

9. The method according to claim 8, wherein the dividing the test samples into corresponding target nodes according to the cluster center point comprises:

10. The method according to claim 6, wherein calculating, by the node and the master node, a predicted value of the satellite positioning signal strength of the test sample according to the parameters of the training subsample, the supporting set and the initialized particle swarm based on the PSO-GPR model comprises:

11. A satellite positioning signal strength prediction apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 10 when executing the computer program.

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