Position estimation method in ship cabin based on Bluetooth technology

Abstract

The present invention provides a method for estimating a position in a ship cabin based on Bluetooth technology, and specifically belongs to the technical field of high-precision positioning services in ship cabins. The present invention performs weighted averaging fusion on an AoA angle estimation result of direct data collection and an AoA angle estimation result calculated by a multivariate sensor to obtain a better AoA angle estimation result, and then fuses the obtained better AoA angle estimation result with a result of RSSI after Kalman filtering to output the final position information. The method for positioning in a ship cabin of the present invention improves the accuracy of positioning in a ship cabin, and is of great significance for ensuring the safety of personnel on board and improving the level of ship operation services.

Description

Technical Field

[0001] The invention belongs to the technical field of high-precision position service in a ship cabin, and in particular relates to a method for estimating a position in a ship cabin based on Bluetooth technology. Background Art

[0002] With the continuous development of my country's shipbuilding industry, high-precision positioning services in ship cabins are not only an important support for ship safety and community operations, but also an important part of realizing ship intelligence. However, unlike conventional building environments, ship cabins have complex spatial layouts and strong signal interference from metal structures, which leads to a decrease in the accuracy of conventional position estimation methods. Therefore, the present invention aims to study high-precision positioning technology suitable for ship cabins, which is of great significance for ensuring passengers' onboard navigation service experience and improving the safety of cruise operations.

[0003] In the existing technology, radio frequency identification or ultra-wideband technology is used for indoor positioning. Although radio frequency identification technology has the advantages of small size, low cost and easy deployment in actual application, RFID is susceptible to interference and has no communication capability. It is not suitable for the technical scenario of indoor positioning service in ship cabins. Although ultra-wideband technology has relatively high positioning accuracy, its application cost is relatively high. It is also not suitable for the technical scenario of indoor positioning service in ship cabins. Summary of the invention

[0004] In order to meet the requirements of high-precision and low-cost positioning in ship cabins, the present invention provides a method for estimating position in ship cabins based on Bluetooth technology. The method can be used in ship indoor scenes to provide more accurate indoor position information.

[0005] In a first aspect of the present invention, the present invention provides a method for estimating a position in a ship cabin based on Bluetooth technology, comprising:

[0006] Step S1, calculate Bluetooth AoA angle estimation based on spatial power spectrum ,

[0007] Step S2, calculate the Bluetooth AoA angle estimation based on the two-stage multi-sensor ;

[0008] Step S3: Bluetooth AoA angle estimation result based on spatial power spectrum Bluetooth AoA Angle Estimation Based on Two-Level Multi-Perceptron Perform adaptive weighted fusion to obtain better angle estimation ;

[0009] Step S4, calculating the positioning coordinates based on the Kalman filter fusion signal strength and the AoA angle.

[0010] Further, step S1 includes:

[0011] Step S11, using the sensor direct data acquisition method to calculate the arrival angle and spatial power spectrum ;

[0012] Step S12, design a joint classifier to classify the arrival angle obtained in step S11 Perform optimization calculations to obtain Bluetooth AoA angle estimation results based on spatial power spectrum .

[0013] Furthermore, step S11 includes:

[0014] Calculate the angle of arrival using direct sensor data acquisition and spatial power spectrum ;

[0015] The calculation method of PDDA spatial power spectrum is:

[0016]

[0017] in, is a scalar value that tends to be infinitesimal and is added to avoid possible singularities. is to construct a power spectrum that obtains a zero point in the direction of the input signal, The calculation method is:

[0018]

[0019] in, represents the constructed spatial spectrum, express The maximum value point, The calculation method is:

[0020]

[0021] in, is a uniform linear array with M elements, it has an M×1-dimensional steering vector, It can be expressed as:

[0022]

[0023] in, Indicates the antenna spacing distance, middle The calculation method is:

[0024]

[0025] in, The calculation method is:

[0026]

[0027] in, and The received signal matrix The two sub-matrices are divided into It is expressed as:

[0028]

[0029] It is expressed as:

[0030]

[0031] It is expressed as:

[0032]

[0033] Right now:

[0034]

[0035] in, is the transmission signal from the signal source, is additive Gaussian white noise with a mean of 0 and a variance of ;

[0036] Angle of arrival is the spatial power spectrum The angle value corresponding to the peak.

[0037] Further, step S12 includes:

[0038] Designing a joint classifier , the specific formula is as follows:

[0039]

[0040] in, are the K position intervals with the most members in the sample data, is the standard deviation of the arrival angle, is the standard deviation of the spatial power spectrum;

[0041] Based on the joint classifier Calculating Bluetooth AoA angle estimate :

[0042]

[0043] in, is the class index corresponding to the direct path from the collection, Represents the voting parameter. If categories belong to the classifier The output of the joint classifier is "1", otherwise the output is "null". The output adopts the absolute majority voting process strategy. Classification from the joint classifier If the total number of votes obtained exceeds half, then The average value of AoA in the class is output as the direction of the direct path of the signal source, that is, the Bluetooth AoA angle estimation result ; otherwise, in the case where no class receives a majority of votes, it will output “no direct path”.

[0044] Further, step S2 includes:

[0045] Step S21, pre-processing the Bluetooth receiver sampling data; first remove the absolute phase information in the IQ value, retain the phase information difference with stronger correlation, and the phase information difference The calculation is as follows:

[0046]

[0047] in, Indicates the phase corresponding to when the signal reaches the first array element; Indicates that the signal reaches The phase corresponding to the array element.

[0048] Step S22, constructing a dual-stage multi-sensor Bluetooth AoA angle estimation model;

[0049] Step S23, using data under interference such as weak attenuation of multipath signals and non-line-of-sight caused by dynamic random disturbances in a ship cabin environment to perform model training;

[0050] Step S24, using the trained two-stage multivariate perceptron model to perform Bluetooth AoA angle estimation calculation to obtain the Bluetooth AoA angle calculation result .

[0051] Further, step S22 includes:

[0052] The model first samples data from each reference node, and then obtains a feature vector with a model input length of 10 after preprocessing. Each vector consists of 4 pairs of standardized IQ values, standardized received signal energy values, and the square sum of the IQ data amplitudes at the sampling position points as reference feature values, which are called signal vectors. Each signal vector is fed to a 4-layer neural network MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer. The neural network generates a 5-dimensional potential representation, and then connects the 5 potential representations with the RSSI and feeds them to a 4-layer channel fusion MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer, to obtain an angle estimate.

[0053] Furthermore, step S3 specifically includes:

[0054] Step S31, calculating the mean square error of Bluetooth AoA angle estimation based on spatial power spectrum calculation And the mean square error of Bluetooth AoA angle estimation based on two-stage multi-source sensor calculation ;

[0055] Step S32, calculating weighted weights; for The weight of for The weight of , The calculation method is:

[0056]

[0057] Where i is 1 or 2. , satisfy The calculation method is:

[0058]

[0059] Step S33, calculate the fused Bluetooth AoA angle estimation value :

[0060] .

[0061] Further, step S4 includes:

[0062] Step S41, combining distance measurement and direction measurement to obtain location information, specifically includes:

[0063] First, the RSSI and AoA measurement models can be approximated using the following formulas:

[0064]

[0065]

[0066] in,

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073] Secondly, the RSSI measurement model and AoA measurement model are expressed in vector form as follows:

[0074]

[0075] in,

[0076] Finally, the position information can be obtained by calculating the equation expressed in the above vector form through the least squares method:

[0077]

[0078] Step S42, Kalman filtering is performed on the position information obtained by combining distance measurement and direction measurement. The specific formula is as follows:

[0079]

[0080]

[0081] Update step, mean and covariance can be calculated by the following formula:

[0082]

[0083]

[0084] in is the Kalman filter, Moment gain.

[0085] In a second aspect of the present invention, the present invention provides a computer device comprising a memory, a processor, and program instructions stored in the memory and executable by the processor, wherein the processor executes the program instructions to implement any one of the methods in the first aspect.

[0086] In a third aspect of the present invention, the present invention provides a computer-readable storage medium storing a computer program, characterized in that when the computer program is executed by a processor, any one of the methods in the first aspect is implemented.

[0087] The present invention provides a method for estimating a position in a ship cabin based on Bluetooth technology, and specifically belongs to the technical field of high-precision positioning services in ship cabins. The present invention performs weighted averaging fusion on an AoA angle estimation result of direct data collection and an AoA angle estimation result calculated by a multivariate sensor to obtain a better AoA angle estimation result, and then fuses the obtained better AoA angle estimation result with a result of RSSI after Kalman filtering to output the final position information. The method for positioning in a ship cabin of the present invention improves the accuracy of positioning in a ship cabin, and is of great significance for ensuring the safety of personnel on board and improving the level of ship operation services. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 It is a schematic diagram of the overall process of the method for estimating the position in a ship cabin based on Bluetooth technology;

[0089] Figure 2 The position estimated by the method of the present invention is compared with the actual position;

[0090] Figure 3 The positioning error results of the method of the present invention and the existing positioning methods are compared. DETAILED DESCRIPTION

[0091] The following will be combined with the drawings in the embodiments of the present application to further clearly and completely describe the technical solutions in the embodiments of the present application. It should be noted that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0092] In order to make the invention purpose, technical scheme and advantages of this application clearer, the technical scheme of the present invention is described clearly and completely in conjunction with the accompanying drawings and specific implementation schemes. It should be pointed out that for ordinary technicians in this field, several changes and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of this application.

[0093] like Figure 1 As shown, in the first embodiment of the present invention, a method for estimating the position in a ship cabin based on Bluetooth technology is provided, comprising step S1, estimating the Bluetooth AoA angle based on the spatial power spectrum ,

[0094] Step S2, calculate the Bluetooth AoA angle estimation based on the two-stage multi-sensor ;

[0095] Step S3: Bluetooth AoA angle estimation result based on spatial power spectrum Bluetooth AoA Angle Estimation Based on Two-Level Multi-Perceptron Perform adaptive weighted fusion to obtain better angle estimation ;

[0096] Step S4, calculating the positioning coordinates based on the Kalman filter fusion signal strength and the AoA angle.

[0097] Further, step S1 includes:

[0098] Step S11, using the sensor direct data acquisition method to calculate the arrival angle and spatial power spectrum ;

[0099] Step S12, design a joint classifier to classify the arrival angle obtained in step S11 Perform optimization calculations to obtain Bluetooth AoA angle estimation results based on spatial power spectrum .

[0100] Furthermore, step S11 includes:

[0101] Calculate the angle of arrival using direct sensor data acquisition and spatial power spectrum ;

[0102] The calculation method of PDDA spatial power spectrum is:

[0103] (1)

[0104] in, is a scalar value that tends to be infinitesimal and is added to avoid possible singularities. is to construct a power spectrum that obtains a zero point in the direction of the input signal, The calculation method is:

[0105] (2)

[0106] in, represents the constructed spatial spectrum, express The maximum value point, The calculation method is:

[0107] (3)

[0108] in, is a uniform linear array with M elements, it has an M×1-dimensional steering vector, It can be expressed as:

[0109] (4)

[0110] in, Indicates the antenna spacing distance, middle The calculation method is:

[0111] (5)

[0112] in, The calculation method is:

[0113] (6)

[0114] in, and The received signal matrix The two sub-matrices are divided into It is expressed as:

[0115] (7)

[0116] It is expressed as:

[0117] (8)

[0118] It is expressed as:

[0119] (9)

[0120] Right now:

[0121] (10)

[0122] in, is the transmission signal from the signal source, is additive Gaussian white noise with a mean of 0 and a variance of ;

[0123] Angle of arrival is the spatial power spectrum The angle value corresponding to the peak.

[0124] Further, step S12 includes:

[0125] Designing a joint classifier , the specific formula is as follows:

[0126] (11)

[0127] Among them, K is the K position intervals with the most members in the sample data, is the standard deviation of the arrival angle, is the standard deviation of the spatial power spectrum;

[0128] Based on the joint classifier Calculating Bluetooth AoA angle estimate :

[0129] (12)

[0130] in, From the joint classifier The category index corresponding to the direct path in , Represents the voting parameter. If categories belong to the classifier The output of the joint classifier is "1", otherwise the output is "null". The output adopts the absolute majority voting process strategy. Classification from the joint classifier If the total number of votes obtained exceeds half, then The average value of the arrival angle (AoA) in the class is output as the azimuth of the direct path of the signal source, that is, the Bluetooth AoA angle estimation result ; otherwise, in the case where no class receives a majority of votes, it will output “no direct path”.

[0131] Further, step S2 includes:

[0132] Step S21, pre-processing the Bluetooth receiver sampling data; first remove the absolute phase information in the IQ value, retain the phase information difference with stronger correlation, and the phase information difference The calculation is as follows:

[0133] (13)

[0134] in, Indicates the phase corresponding to when the signal reaches the first array element; Indicates that the signal reaches The phase corresponding to the array element.

[0135] Step S22, constructing a dual-stage multi-sensor Bluetooth AoA angle estimation model;

[0136] Step S23, using data under interference such as weak attenuation of multipath signals and non-line-of-sight caused by dynamic random disturbances in a ship cabin environment to perform model training;

[0137] Step S24, using the trained two-stage multivariate perceptron model to perform Bluetooth AoA angle estimation calculation to obtain the Bluetooth AoA angle calculation result .

[0138] Further, step S22 includes:

[0139] The model first samples data from each reference node, and then obtains a feature vector with a model input length of 10 after preprocessing. Each vector consists of 4 pairs of standardized IQ values, standardized received signal energy values, and the square sum of the IQ data amplitudes at the sampling position points as reference feature values, which are called signal vectors. Each signal vector is fed to a 4-layer neural network MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer. The neural network generates a 5-dimensional potential representation, and then connects the 5 potential representations with the RSSI and feeds them to a 4-layer channel fusion MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer, to obtain an angle estimate.

[0140] Furthermore, step S3 specifically includes:

[0141] Step S31, calculating the mean square error of Bluetooth AoA angle estimation based on spatial power spectrum calculation And the mean square error of Bluetooth AoA angle estimation based on two-stage multi-source sensor calculation ;

[0142] Step S32, calculating weighted weights; for The weight of for The weight of , The calculation method is:

[0143] (14)

[0144] in, Choose 1 or 2. , satisfy The calculation method is:

[0145] (15)

[0146] Step S33, calculate the fused Bluetooth AoA angle estimation value :

[0147] (16).

[0148] Further, step S4 includes:

[0149] Step S41, combining distance measurement and direction measurement to obtain location information, specifically includes:

[0150] First, the RSSI and AoA measurement models can be approximated using the following two formulas:

[0151] (17)

[0152] (18)

[0153] in,

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160] Secondly, the RSSI measurement model and AoA measurement model are expressed in vector form as follows:

[0161] (19)

[0162] in,

[0163] Finally, the position information can be obtained by calculating the above formula (19) using the least squares method:

[0164] (20);

[0165] Step S42, Kalman filtering is performed on the position information obtained by combining distance measurement and direction measurement. The specific formula is as follows:

[0166] (twenty one)

[0167] (twenty two)

[0168] Update step, mean and covariance can be calculated by the following formula:

[0169] (twenty three)

[0170] (twenty four)

[0171] in is the Kalman filter, Moment gain.

[0172] In a second embodiment of the present invention, the present invention provides a computer device, comprising a memory, a processor, and program instructions stored in the memory and executable by the processor, wherein the processor executes the program instructions to implement any one of the methods in the first aspect.

[0173] In a third embodiment of the present invention, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program implements any one of the methods in the first aspect when executed by a processor.

[0174] Attached Figure 2 What is shown is the comparison between the position estimated by the method of the present invention and the actual position, the circle is the actual position, and the triangulated point is the position estimation result; through numerical calculation and analysis of the algorithm performance, it can be obtained that in the experiment of fixed-point position estimation, the maximum Euclidean distance position error is about 37cm, and the average Euclidean distance position error is about 23cm, which verifies that the position accuracy requirements are met.

[0175] Attached Figure 3 The figure shows the positioning error comparison between the method of the present invention and the existing positioning method, specifically the mean square error of the method of the present invention and the mean square error of PDDA. From the width of the interquartile range shown in the figure, it can be seen that the error distribution of the positioning method proposed by the present invention is more concentrated. It can be seen that the method proposed by the present invention can effectively improve the Bluetooth AoA positioning accuracy in the ship cabin environment, and compared with the conventional positioning method using PDDA, the average accuracy is improved by about 41.7%.

[0176] In summary, the present invention provides a method for estimating the position in a ship cabin based on Bluetooth technology.

[0177] Providing more accurate indoor positioning services for the indoor environment of ships with complex spatial layouts and strong signal interference from metal structures is of great significance to improving shipboard safety and community intelligent operations and realizing ship intelligence.

Claims

1. A method for estimating position in a ship cabin based on Bluetooth technology, characterized in that: include: Step S1, calculate Bluetooth AoA angle estimation based on spatial power spectrum , Step S2, calculate the Bluetooth AoA angle estimation based on the two-stage multi-sensor ; Step S2 comprises: Step S21, pre-processing the Bluetooth receiver sampling data; first remove the absolute phase information in the IQ value, retain the phase information difference with stronger correlation, and the phase information difference The calculation is as follows: （13）； in, Indicates the phase corresponding to when the signal reaches the first array element; Indicates that the signal reaches The corresponding phase when there are array elements; Step S22, constructing a dual-stage multi-sensor Bluetooth AoA angle estimation model; Step S23, using data of weak attenuation of multipath signals and non-line-of-sight interference caused by dynamic random disturbance in a ship cabin environment to perform model training; Step S24, using the trained two-stage multivariate perceptron model to perform Bluetooth AoA angle estimation calculation to obtain the Bluetooth AoA angle calculation result ; The step S22 comprises: The model first samples data from each reference node, and then obtains a feature vector with a model input length of 10 after preprocessing. Each vector consists of 4 pairs of standardized IQ values, standardized received signal energy values, and the square sum of IQ data amplitudes at the sampling position points as reference feature values, which are called signal vectors. Each signal vector is fed to a 4-layer neural network MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer. The neural network generates a 5-dimensional potential representation, and then connects the 5 potential representations with RSSI and feeds them to a 4-layer channel fusion MLP, namely the input layer, the first hidden layer, the second hidden layer, and the output layer, to obtain an angle estimate. Step S3: Bluetooth AoA angle estimation result based on spatial power spectrum Bluetooth AoA Angle Estimation Based on Two-Level Multi-Perceptron Perform adaptive weighted fusion to obtain better angle estimation ; Step S4, calculating the positioning coordinates based on the Kalman filter fusion signal strength and the AoA angle.

2. According to the method for estimating the position in a ship cabin based on Bluetooth technology in claim 1, it is characterized in that: The step S1 comprises: Step S11, using the sensor direct data acquisition method to calculate the arrival angle and spatial power spectrum ; Step S12, design a joint classifier to classify the arrival angle obtained in step S11 Perform optimization calculations to obtain Bluetooth AoA angle estimation results based on spatial power spectrum .

3. According to the method for estimating position in a ship cabin based on Bluetooth technology in claim 2, it is characterized in that: The step S11 comprises: Calculate the angle of arrival using direct sensor data acquisition and spatial power spectrum ; The calculation method of PDDA spatial power spectrum is: （1）； in, is a scalar value that tends to be infinitesimal and is added to avoid possible singularities. is to construct a power spectrum that obtains a zero point in the direction of the input signal, The calculation method is: （2）； in, represents the constructed spatial spectrum, ω represents The maximum value point, The calculation method is: （3）； in, is a uniform linear array with M elements, it has an M×1-dimensional steering vector, It can be expressed as: （4）； in, Indicates the antenna spacing distance, middle The calculation method is: （5）； in, The calculation method is: （6）； in, and The received signal matrix The two sub-matrices are divided into It is expressed as: （7）； It is expressed as: （8）； It is expressed as: （9）； Right now: （10）； in, is the transmission signal from the signal source, is additive Gaussian white noise with a mean of 0 and a variance of ; Angle of arrival is the spatial power spectrum The angle value corresponding to the peak.

4. According to the method for estimating position in a ship cabin based on Bluetooth technology in claim 3, it is characterized in that: The step S12 comprises: Designing a joint classifier , the specific formula is as follows: （11）； Among them, K is the K position intervals with the most members in the sample data, is the standard deviation of the arrival angle, is the standard deviation of the spatial power spectrum; Based on the joint classifier Calculating Bluetooth AoA angle estimate : （12）； in, From the joint classifier The category index corresponding to the direct path in , Represents the voting parameter. If categories belong to the classifier The output of the joint classifier is "1", otherwise the output is "null". The output adopts the absolute majority voting process strategy. Classification from the joint classifier If the total number of votes obtained exceeds half, then The average value of AoA in the class is output as the direction of the direct path of the signal source, that is, the Bluetooth AoA angle estimation result ; otherwise, in the case where no class gets a majority of votes, it will output "no direct path".

5. According to the method for estimating position in a ship cabin based on Bluetooth technology in claim 1, it is characterized in that: The step S3 specifically includes: Step S31, calculating the mean square error of Bluetooth AoA angle estimation based on spatial power spectrum calculation And the mean square error of Bluetooth AoA angle estimation based on two-stage multi-source sensor calculation ; Step S32, calculating weighted weights; for The weight of for The weight of , The calculation method is: （14）； in, Choose 1 or 2. , satisfy The calculation method is: （15）； Step S33, calculate the fused Bluetooth AoA angle estimation value : （16）。 6. According to the method for estimating position in a ship cabin based on Bluetooth technology in claim 1, it is characterized in that: The step S4 comprises: Step S41, combining distance measurement and direction measurement to obtain location information, specifically includes: First, the RSSI and AoA measurement models can be approximated using the following two equations: （17）； （18）； in, ； ； ； ； ； ； Secondly, the RSSI measurement model and AoA measurement model are expressed in vector form as follows: （19）； in, , Finally, the position information can be obtained by calculating the above formula (19) using the least squares method: （20）； Step S42, Kalman filtering is performed on the position information obtained by combining distance measurement and direction measurement. The specific formula is as follows: （21）； （22）； Update step, mean and covariance can be calculated by the following formula: （23）； （24）； in is the Kalman filter, Moment gain.

7. A computer device comprising a memory, a processor and program instructions stored in the memory and executable by the processor, characterized in that: The processor executes the program instructions to implement any one of the methods of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images