A smart bed positioning system and method based on RFID

Abstract

The present invention relates to an intelligent hospital bed positioning system and method based on RFID, which obtains the main feature data set of RFID tag signal strength by inputting a time series RFID tag signal strength data set into an autoencoder; inputting the main feature data set of RFID tag signal strength into a preset Bayesian network, and outputting the probability value of each distance parameter of the RFID tag; selecting the distance parameter with the largest probability value, and identifying the distance range contained in the distance parameter; taking the position of the RFID reader as the center of the circle, drawing circles with the upper and lower limits of the numerical value of the distance range as the radius, and calculating the area difference between the two circles; by moving the position of the RFID reader, analyzing the change of the area difference between the two circles, to determine the direction of the RFID tag. The present invention solves the problem that the RFID signal strength is unstable and its distance and direction are difficult to be determined by the signal strength, resulting in poor positioning effect.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical equipment and relates to an intelligent hospital bed positioning system and method based on RFID. Background Art

[0002] There are many wards in the hospital inpatient department, and a large number of beds are provided. For patients and doctors who are not familiar with the hospital environment, they often get lost and cannot find the corresponding bed location. RFID, or radio frequency identification, is a non-contact automatic identification technology that can achieve management functions such as fast reading and writing, long-term tracking, and has a very good development prospect in the field of intelligent identification. Therefore, applying RFID to hospital beds can help doctors and patients quickly locate beds and improve medical work efficiency.

[0003] The basic principle of RFID is that the RFID reader obtains the signal strength of the RFID tag to determine the location of the object to which the RFID tag is attached. In actual applications, the signal strength of the RFID tag is not stable and will continue to change, so a special algorithm is needed to estimate the exact location of the RFID tag. In addition, the RFID tag signal strength is difficult to be used to determine the distance and direction of the RFID tag, resulting in poor positioning results and other problems. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides an RFID-based smart bed positioning system and method.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] In the first aspect, the present application provides an RFID-based smart bed positioning system, including a signal search module, a location analysis module and a guidance module, wherein:

[0007] The signal search module has a built-in RFID reader / writer for searching for the RFID tag signal attached to the bed to be located;

[0008] The position analysis module is connected to the signal search module and is used to determine the distance and direction between the RFID reader and the bed according to the RFID tag signal strength;

[0009] The guiding module is connected to the position analyzing module and is used to guide the patient or doctor to find the bed to be located according to the distance and direction between the RFID reader and the bed.

[0010] Furthermore, in the position analysis module, determining the distance and direction between the RFID reader and the hospital bed according to the RFID tag signal strength includes the following steps:

[0011] S1, taking t seconds as the time step, obtain a time series of RFID tag signal strength data sets;

[0012] S2. Input the RFID tag signal strength data set into the autoencoder to obtain the main feature data set of the RFID tag signal strength;

[0013] S3, inputting the main feature data set of the RFID tag signal strength into a preset Bayesian network, and outputting the probability value of each distance parameter of the RFID tag;

[0014] S4, selecting a distance parameter with a maximum probability value, and identifying a distance range included in the distance parameter;

[0015] S5, taking the position of the RFID reader as the center of the circle, drawing a circle with the upper and lower limits of the distance range as the radius, and calculating the area difference between the two circles;

[0016] S6. By moving the position of the RFID reader, the change in the difference between the areas of the two circles is analyzed to determine the direction of the RFID tag.

[0017] Furthermore, in step S2, the step of inputting the RFID tag signal strength data set into the autoencoder to obtain the main feature data set of the RFID tag signal strength includes the following steps:

[0018] S21, using the encoder of the autoencoder to compress the RFID tag signal strength data set, and the calculation formula is: ,

[0019] Where: Z is the latent variable, which is the input variable after compression processing; X is the input variable, that is, the RFID tag signal strength data set; and Represent the weights and biases of the encoding part of the model respectively; Represents the Sigmoid activation function of the encoder;

[0020] S22. The latent variables are processed by the decoder of the autoencoder to reconstruct the original data. The calculation formula is as follows: ,

[0021] Where: is the output data, which is the reconstructed original data; Z is the latent variable, which is the input variable after compression processing; and Represent the weights and biases of the decoding part of the model respectively; Represents the Sigmoid activation function of the decoder.

[0022] Furthermore, in step S3, the preset Bayesian network is constructed as follows:

[0023] S31, data collection: collecting RFID tag signal strength data at different distances between the RFID reader and the tag, and obtaining RFID tag distance and RFID tag signal strength data sets for building the model;

[0024] S32, model structure determination: determine the structure of the model based on the collected data variables;

[0025] S33, parameter estimation: using the maximum likelihood estimation method to determine the parameters of the probability distribution of the RFID tag distance and RFID tag signal strength data variables, and constructing a conditional probability table based on the parameter estimation results;

[0026] S34, model structure learning: use the structure learning algorithm to further adjust and optimize the network structure to obtain a better network structure;

[0027] S35. Model validation: Use known prediction data sets to verify the accuracy of the model.

[0028] Furthermore, in step S32, the model structure uses the RFID tag signal strength as a parent node and the RFID tag distance as a child node, and connects the parent node and the child node.

[0029] Furthermore, in step S33, the maximum likelihood estimation method is used to determine the parameters of the probability distribution of the RFID tag distance and RFID tag signal strength data variables, and a conditional probability table is constructed according to the parameter estimation results, including the following steps:

[0030] S331, constructing a likelihood function: constructing a likelihood function according to the distribution of the observed data and the parameters of the model, wherein the likelihood function is recorded as: ,

[0031] Where L(θ | D) is the likelihood function, which indicates the probability that the parameter of the distribution function is θ in a given observation sample D; P(D | θ) is the conditional probability function, which indicates the probability that the observation result is D when the parameter θ is known;

[0032] S332, Maximizing the likelihood function: By maximizing the likelihood function, find the maximum parameter value that makes the observed data most likely , recorded as: ;

[0033] S333, estimated parameters: the parameter value obtained by maximizing the likelihood function is the maximum likelihood estimate of the parameter;

[0034] S334, constructing a conditional probability table: constructing a conditional probability table for each variable based on the parameter values ​​obtained by parameter estimation. For each variable, given its parent node, calculate its joint probability at multiple parent nodes. The formula is:

[0035] ,

[0036] Where P(U) is a set of nodes U = B1, B2, ..., B n The joint probability distribution of Pa(B i ) is the node B in the Bayesian network i The parent node set of P(B i ) represents the target node B i The prior probability of .

[0037] Furthermore, in step S34, the structure learning algorithm is the Bayesian Information Criterion BIC. By enumerating all network structures, the network structure with the smallest BIC value is selected as the optimal model structure. The calculation formula of the Bayesian Information Criterion BIC is:

[0038] ,

[0039] Where L is the maximum likelihood function estimate of the model; k is the number of estimated free parameters; and n is the number of samples.

[0040] Furthermore, in step S6, the position of the RFID reader is moved to analyze the change in the difference between the two circles to determine the direction of the RFID tag, specifically:

[0041] When the difference between the two circle areas becomes larger, it means that the RFID reader is in the direction of starting to move away from the RFID tag;

[0042] When the difference between the areas of the two circles becomes smaller, it means that the RFID reader is in the direction of approaching the RFID tag.

[0043] In a second aspect, the present application provides an RFID-based smart bed positioning method, which is applied to an RFID-based smart bed positioning system as described above, and includes the following steps:

[0044] Using t seconds as the time step, obtain a data set of RFID tag signal strength over a period of time;

[0045] The RFID tag signal strength dataset is input into the autoencoder to obtain the main feature dataset of the RFID tag signal strength;

[0046] Input the main feature data set of RFID tag signal strength into the preset Bayesian network and output the probability value of each distance parameter of the RFID tag;

[0047] Select the distance parameter with the largest probability value, and identify the distance range included in the distance parameter;

[0048] Taking the position of the RFID reader as the center of the circle, draw circles with the upper and lower limits of the distance range as the radius, and calculate the area difference between the two circles;

[0049] By moving the position of the RFID reader and writer and analyzing the change in the area difference between the two circles, the direction of the RFID tag can be determined.

[0050] Beneficial effects of the present invention:

[0051] By inputting a time series of RFID tag signal strength data sets into an autoencoder, the main feature data sets of RFID tag signal strength are obtained; the main feature data sets of RFID tag signal strength are input into a preset Bayesian network, and the probability values ​​of each distance parameter of the RFID tag are output; the distance parameter with the largest probability value is selected, and the distance range contained in the distance parameter is identified; with the position of the RFID reader as the center of the circle, circles are drawn with the upper and lower limits of the numerical value of the distance range as the radius, and the area difference between the two circles is calculated; by moving the position of the RFID reader, the change of the area difference between the two circles is analyzed to determine the direction of the RFID tag. The present invention solves the problem that the signal strength of the RFID tag fluctuates and the distance and direction of the RFID tag are difficult to be determined by the signal strength, resulting in inaccurate positioning and poor effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0053] Figure 1 This is a structural diagram of an RFID-based smart bed positioning system in the present invention. DETAILED DESCRIPTION

[0054] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0055] See also Figure 1 In the first aspect, the present application provides an RFID-based smart bed positioning system, including a signal search module, a location analysis module and a guidance module, wherein:

[0056] The signal search module has a built-in RFID reader / writer for searching for the RFID tag signal attached to the bed to be located;

[0057] The position analysis module is connected to the signal search module and is used to determine the distance and direction between the RFID reader and the bed according to the RFID tag signal strength;

[0058] The guiding module is connected to the position analyzing module and is used to guide the patient or doctor to find the bed to be located according to the distance and direction between the RFID reader and the bed.

[0059] Furthermore, in the position analysis module, determining the distance and direction between the RFID reader and the hospital bed according to the RFID tag signal strength includes the following steps:

[0060] S1, taking t seconds as the time step, obtain a time series of RFID tag signal strength data sets;

[0061] S2. Input the RFID tag signal strength data set into the autoencoder to obtain the main feature data set of the RFID tag signal strength;

[0062] S3, inputting the main feature data set of the RFID tag signal strength into a preset Bayesian network, and outputting the probability value of each distance parameter of the RFID tag;

[0063] S4, selecting a distance parameter with a maximum probability value, and identifying a distance range included in the distance parameter;

[0064] S5, taking the position of the RFID reader as the center of the circle, drawing a circle with the upper and lower limits of the distance range as the radius, and calculating the area difference between the two circles;

[0065] S6. By moving the position of the RFID reader, the change in the difference between the areas of the two circles is analyzed to determine the direction of the RFID tag.

[0066] It should be noted that due to various interference factors, the signal strength of the searched RFID tag may fluctuate. In order to reduce the impact of interference factors, this embodiment obtains a time series of RFID tag signal strength data sets, inputs it into the autoencoder to reduce the noise of the data set, and obtains the main features of the data set. The autoencoder is an unsupervised learning algorithm that learns the compact representation of input data through the structure of the encoder and decoder. It can be used for tasks such as data dimensionality reduction, feature learning, data repair, data generation, and feature extraction. By mapping the input data to the potential representation space through the encoder and then mapping the potential representation back to the original data space through the decoder, the autoencoder can learn the important features of the input data and use it in various practical applications.

[0067] Furthermore, in step S2, the step of inputting the RFID tag signal strength data set into the autoencoder to obtain the main feature data set of the RFID tag signal strength includes the following steps:

[0068] S21, using the encoder of the autoencoder to compress the RFID tag signal strength data set, and the calculation formula is: ,

[0069] Where: Z is the latent variable, which is the input variable after compression processing; X is the input variable, that is, the RFID tag signal strength data set; and Represent the weights and biases of the encoding part of the model respectively; Represents the Sigmoid activation function of the encoder;

[0070] S22. The latent variables are processed by the decoder of the autoencoder to reconstruct the original data. The calculation formula is as follows: , where: is the output data, which is the reconstructed original data; Z is the latent variable, which is the input variable after compression processing; and Represent the weights and biases of the decoding part of the model respectively; Represents the Sigmoid activation function of the decoder.

[0071] In this embodiment, a Bayesian network model is constructed by using the correlation between the received RFID tag signal strength and the distance. In this process, the distance is parameterized to obtain distance parameters containing different distance ranges. This parameterization process can include a variety of processing methods. The distance parameters can be divided based on empirical judgment; cluster analysis methods can be used to divide the distance parameters, such as K-means clustering, etc.; the maximum likelihood estimation method can also be used to determine the parameters of the probability distribution of each variable in the Bayesian network model. The constructed Bayesian network is applied to the monitoring of RFID tag signals, and the distance range of the RFID tag can be determined according to the RFID tag signal strength.

[0072] Furthermore, in step S3, the preset Bayesian network is constructed as follows:

[0073] S31, data collection: collecting RFID tag signal strength data at different distances between the RFID reader and the tag, and obtaining RFID tag distance and RFID tag signal strength data sets for building the model;

[0074] S32, model structure determination: determine the structure of the model based on the collected data variables;

[0075] S33, parameter estimation: using the maximum likelihood estimation method to determine the parameters of the probability distribution of the RFID tag distance and RFID tag signal strength data variables, and constructing a conditional probability table based on the parameter estimation results;

[0076] S34, model structure learning: use the structure learning algorithm to further adjust and optimize the network structure to obtain a better network structure;

[0077] S35. Model validation: Use known prediction data sets to verify the accuracy of the model.

[0078] Furthermore, in step S32, the model structure uses the RFID tag signal strength as a parent node and the RFID tag distance as a child node, and connects the parent node and the child node.

[0079] Furthermore, in step S33, the maximum likelihood estimation method is used to determine the parameters of the probability distribution of the RFID tag distance and RFID tag signal strength data variables, and a conditional probability table is constructed according to the parameter estimation results, including the following steps:

[0080] S331, constructing a likelihood function: constructing a likelihood function according to the distribution of the observed data and the parameters of the model, wherein the likelihood function is recorded as: ,

[0081] Where L(θ | D) is the likelihood function, which indicates the probability that the parameter of the distribution function is θ in a given observation sample D; P(D | θ) is the conditional probability function, which indicates the probability that the observation result is D when the parameter θ is known;

[0082] S332, Maximizing the likelihood function: By maximizing the likelihood function, find the maximum parameter value that makes the observed data most likely , recorded as: ;

[0083] S333, estimated parameters: the parameter value obtained by maximizing the likelihood function is the maximum likelihood estimate of the parameter;

[0084] S334, constructing a conditional probability table: constructing a conditional probability table for each variable based on the parameter values ​​obtained by parameter estimation. For each variable, given its parent node, calculate its joint probability at multiple parent nodes. The formula is:

[0085] ,

[0086] Where P(U) is a set of nodes U = B1, B2, ..., B n The joint probability distribution of Pa(B i ) is the node B in the Bayesian network i The parent node set of P(B i ) represents the target node B i The prior probability of .

[0087] Furthermore, in step S34, the structure learning algorithm is the Bayesian Information Criterion BIC. By enumerating all network structures, the network structure with the smallest BIC value is selected as the optimal model structure. The calculation formula of the Bayesian Information Criterion BIC is:

[0088] ,

[0089] Where L is the maximum likelihood function estimate of the model; k is the number of estimated free parameters; and n is the number of samples.

[0090] Furthermore, in step S6, the position of the RFID reader is moved to analyze the change in the difference between the two circles to determine the direction of the RFID tag, specifically:

[0091] When the difference between the two circle areas becomes larger, it means that the RFID reader is in the direction of starting to move away from the RFID tag;

[0092] When the difference between the areas of the two circles becomes smaller, it means that the RFID reader is in the direction of approaching the RFID tag.

[0093] In a second aspect, the present application provides an RFID-based smart bed positioning method, which is applied to an RFID-based smart bed positioning system as described above, and includes the following steps:

[0094] Using t seconds as the time step, obtain a data set of RFID tag signal strength over a period of time;

[0095] The RFID tag signal strength dataset is input into the autoencoder to obtain the main feature dataset of the RFID tag signal strength;

[0096] Input the main feature data set of RFID tag signal strength into the preset Bayesian network and output the probability value of each distance parameter of the RFID tag;

[0097] Select the distance parameter with the largest probability value, and identify the distance range included in the distance parameter;

[0098] Taking the position of the RFID reader as the center of the circle, draw circles with the upper and lower limits of the distance range as the radius, and calculate the area difference between the two circles;

[0099] By moving the position of the RFID reader and writer and analyzing the change in the area difference between the two circles, the direction of the RFID tag can be determined.

[0100] Beneficial effects of the present invention:

[0101] By inputting a time series of RFID tag signal strength data sets into an autoencoder, a main feature data set of RFID tag signal strength is obtained; the main feature data set of RFID tag signal strength is input into a preset Bayesian network, and the probability value of each distance parameter of the RFID tag is output; the distance parameter with the largest probability value is selected, and the distance range contained in the distance parameter is identified; with the position of the RFID reader as the center of the circle, a circle is drawn with the upper and lower limits of the numerical value of the distance range as the radius, and the area difference between the two circles is calculated; by moving the position of the RFID reader, the change of the area difference between the two circles is analyzed to determine the direction of the RFID tag. The present invention solves the problem that the signal strength of the RFID tag fluctuates and the distance and direction of the RFID tag are difficult to be determined by the signal strength, resulting in inaccurate positioning and poor effect.

[0102] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An RFID-based smart bed positioning system, characterized in that: It includes a signal search module, a position analysis module and a guidance module, wherein: The signal search module has a built-in RFID reader / writer for searching for the RFID tag signal attached to the bed to be located; The position analysis module is connected to the signal search module and is used to determine the distance and direction between the RFID reader and the bed according to the RFID tag signal strength; The guiding module is connected to the location analysis module and is used to guide the patient or doctor to find the bed to be located according to the distance and direction between the RFID reader and the bed; In the position analysis module, determining the distance and direction between the RFID reader and the hospital bed according to the RFID tag signal strength includes the following steps: S1, taking t seconds as the time step, obtain a time series of RFID tag signal strength data sets; S2. Input the RFID tag signal strength data set into the autoencoder to obtain the main feature data set of the RFID tag signal strength; S3, inputting the main feature data set of the RFID tag signal strength into a preset Bayesian network, and outputting the probability value of each distance parameter of the RFID tag; S4, selecting a distance parameter with a maximum probability value, and identifying a distance range included in the distance parameter; S5, taking the position of the RFID reader as the center of the circle, drawing circles with the upper and lower limits of the distance range as radii, and calculating the area difference between the two circles; S6. By moving the position of the RFID reader, analyzing the change in the area difference between the two circles, so as to determine the direction of the RFID tag, specifically: When the area difference between the two circles becomes larger, it means that the RFID reader is in the direction of starting to move away from the RFID tag; When the area difference between the two circles becomes smaller, it means that the RFID reader is in the direction of approaching the RFID tag; In step S3, the preset Bayesian network is constructed as follows: S31, data collection: collecting RFID tag signal strength data at different distances between the RFID reader and the tag, and obtaining RFID tag distance and RFID tag signal strength data sets for building the model; S32, model structure determination: determine the structure of the model based on the collected data variables; S33, parameter estimation: using the maximum likelihood estimation method to determine the parameters of the probability distribution of the RFID tag distance and RFID tag signal strength data variables, and constructing a conditional probability table based on the parameter estimation results; S34, model structure learning: use the structure learning algorithm to further adjust and optimize the network structure to obtain a better network structure; S35. Model validation: Use known prediction data sets to verify the accuracy of the model.

2. The RFID-based smart bed positioning system according to claim 1, characterized in that: In step S32, the model structure uses the RFID tag signal strength as the parent node and the RFID tag distance as the child node, and connects the parent node and the child node.

3. The RFID-based smart bed positioning system according to claim 1, characterized in that: In step S2, the RFID tag signal strength data set is input into the autoencoder to obtain the main feature data set of the RFID tag signal strength, including the following steps: S21, using the encoder of the autoencoder to compress the RFID tag signal strength data set, and the calculation formula is: ， Where: is a latent variable, which is the input variable after compression processing; is the input variable, i.e., the RFID tag signal strength dataset; and Represent the weights and biases of the encoding part of the model respectively; Represents the Sigmoid activation function of the encoder; S22. The latent variables are processed by the decoder of the autoencoder to reconstruct the original data. The calculation formula is as follows: ， Where: is the output data, which is the reconstructed original data; is a latent variable, which is the input variable after compression processing; and Represent the weights and biases of the decoding part of the model respectively; Represents the Sigmoid activation function of the decoder.

4. The RFID-based smart bed positioning system according to claim 1, characterized in that: In step S34, the structure learning algorithm is the Bayesian Information Criterion BIC. By enumerating all network structures, the network structure with the smallest BIC value is selected as the optimal model structure. The calculation formula of the Bayesian Information Criterion BIC is: ， Where L is the maximum likelihood function estimate of the model; k is the number of free parameters estimated; n is the number of samples.

5. A smart bed positioning method based on RFID, characterized in that: An RFID-based smart bed positioning system as described in any one of claims 1 to 4 comprises the following steps: Using t seconds as the time step, obtain a data set of RFID tag signal strength over a period of time; Input the RFID tag signal strength dataset into the autoencoder to obtain the main feature dataset of the RFID tag signal strength; Input the main feature data set of RFID tag signal strength into the preset Bayesian network and output the probability value of each distance parameter of the RFID tag; Select the distance parameter with the largest probability value, and identify the distance range included in the distance parameter; Taking the position of the RFID reader as the center of the circle, draw circles with the upper and lower limits of the distance range as the radius, and calculate the area difference between the two circles; By moving the position of the RFID reader and writer and analyzing the change in the area difference between the two circles, the direction of the RFID tag can be determined.

Images