A method to improve RFID recognition accuracy based on equidistant center calculation

Abstract

The present invention discloses a method for improving RFID recognition accuracy based on equidistant center calculation, which relates to the field of physical asset inventory. The method for improving RFID recognition accuracy based on equidistant center calculation includes the following steps: obtaining the equidistant center and average distance of the space; preliminarily adjusting the radio frequency power of the RFID based on the average distance of the space; adjusting the radio frequency power of the RFID step by step; using the loss model to calculate the signal strength of RFID recognition, and judging whether assets in other spaces are recognized in the current space. The method for improving RFID recognition accuracy based on equidistant center calculation of the present invention can help enterprises find the most suitable radio frequency strength for the current space before asset inventory, and solve the problem of cross-space scanning assets based on the current space and radio frequency strength.

Description

Technical Field

[0001] The present invention relates to the field of physical asset inventory, and in particular to a method for improving RFID recognition accuracy based on equidistant center calculation. Background Art

[0002] Physical asset inventory is one of the important links in the management of physical assets. Regularly checking the quantity and status of various physical assets in the enterprise can help the enterprise understand the actual situation of the assets, identify problems, and do a good job in asset management.

[0003] Common physical assets include fixed assets such as production equipment, office equipment, and computer equipment, as well as current assets such as inventory. Companies usually set different inventory frequencies according to different asset categories, such as annual inventory of fixed assets and quarterly or monthly inventory of inventory.

[0004] In the physical asset inventory business, the more common method is to use scanning equipment based on the spatial dimension and RFID identification tags. RFID identification tags are a type of tag that uses radio frequency identification technology to identify physical assets. It is composed of a micro wireless radio frequency chip and an antenna, and is small in size and can be directly pasted or sewn on the asset. Each tag has a unique serial number that can uniquely identify the corresponding asset. However, in the process of using the device to scan, how to adjust the radio frequency scanning intensity suitable for the space and mistakenly scan physical assets in other spaces are issues that need to be urgently resolved in the asset inventory process, which plays a decisive role in improving the efficiency and accuracy of asset inventory.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related art, the present invention proposes a method for improving RFID recognition accuracy based on equidistant center calculation to overcome the above technical problems existing in the existing related art.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] A method for improving RFID recognition accuracy based on equidistant center calculation, the method for improving RFID recognition accuracy based on equidistant center calculation comprises the following steps:

[0009] S1. Obtain the equidistant center and average distance of the space.

[0010] S2. Perform preliminary calibration on the RFID radio frequency power based on the average distance in space.

[0011] S3. Adjust the RFID radio frequency power step by step.

[0012] S4. Calculate the signal strength of RFID recognition using the loss model, and determine whether assets in other spaces are recognized in the current space.

[0013] Furthermore, obtaining the equidistant center and the average distance of the space includes the following steps:

[0014] S11, selecting a number of points on the boundary of the space;

[0015] S12, construct the objective function and calculate the equidistant center of the space;

[0016] S13. Calculate the average distance in space using the Euclidean distance formula.

[0017] Furthermore, constructing the objective function and calculating the equidistant center of the space includes the following steps:

[0018] S121, define the objective function;

[0019] Among them, the objective function is:

[0020] F(C)=Σ(d(C,P i ));

[0021] Where F(C) represents the objective function;

[0022] d(C,P i ) indicates the distance between the center point C and each point P i distance;

[0023] S122. Use the gradient descent method to minimize the objective function and obtain the equidistant centers.

[0024] Furthermore, the objective function is minimized using the gradient descent method, and obtaining the equidistant centers includes the following steps:

[0025] S1221, initialize d to a random value;

[0026] S1222. Define a loss function;

[0027] S1223, calculating the gradient of the loss function with respect to d;

[0028] S1224, update d using the gradient descent method;

[0029] The calculation formula for updating is:

[0030]

[0031] In the formula, d represents the distance between the center point C and each point P i distance;

[0032] a represents the learning rate;

[0033] L(d) represents the loss function, represents partial derivative operation;

[0034] S1225. Repeat S1223-S1224 until the preset number of iterations is reached.

[0035] Furthermore, calculating the gradient of the loss function with respect to d includes the following steps:

[0036] Expand the loss function and calculate the distance between the center point C and each point P i The distance is used to find the derivative;

[0037] Sum the obtained derivatives to get the gradient of the loss function with respect to d.

[0038] Furthermore, preliminary calibration of the RFID radio frequency power based on the average distance in space includes the following steps:

[0039] S21, setting the average distance of the space as the identification distance of RFID identification, and obtaining the radio frequency power through the radio frequency power calculation formula;

[0040] The RF power calculation formula is:

[0041]

[0042] S22, adjusting the radio frequency power of the RFID identification device to the calculated radio frequency power.

[0043] Furthermore, the step-by-step adjustment of the RFID radio frequency power includes the following steps:

[0044] S31, assigning the equidistant center of the space to the adjustment distance;

[0045] S32, placing the physical asset with the tag at the adjustment distance, and using the set RFID identification device to identify it;

[0046] S33. If it is not recognized, the RF power is increased by 0.1 and recognized again. If it is recognized, the RF power is reduced by 0.1 and recognized again. If it is recognized this time and not recognized in the next operation, the adjustment is completed.

[0047] Furthermore, using the loss model to calculate the signal strength of RFID recognition and determining whether assets in other spaces are recognized in the current space includes the following steps:

[0048] S41, setting a number of calibration signal strength values ​​for different transmission distances, and collecting sample data corresponding to the transmission distances, averaging and standard deviation of all sample data, and obtaining the standard deviation of the sample data;

[0049] S42, obtaining a threshold value to determine whether assets in other spaces are scanned;

[0050] Among them, the formula of the loss model is:

[0051] RSSI(dBm)=x(dBm)-20*log 10 (y)-20*log 10 (f)+147.55;

[0052] Where RSSI (dBm) is the received signal strength;

[0053] x (dBm) is the RF power;

[0054] y is the transmission distance;

[0055] f is the operating frequency.

[0056] Furthermore, the threshold is twice the standard deviation of the sample data.

[0057] Furthermore, when determining whether assets in other spaces are scanned, if the difference between the actual signal strength value and the calibrated signal strength value is greater than a threshold, it is determined that the current space recognizes assets in other spaces.

[0058] The beneficial effects of the present invention are:

[0059] (1) The method for improving RFID recognition accuracy based on equidistant center calculation of the present invention can help enterprises find the most suitable radio frequency strength for the current space before asset inventory and solve the problem of cross-space asset scanning based on the current space and radio frequency strength.

[0060] (2) The present invention utilizes the equidistant center of space as the farthest distance for RFID identification, which can ensure that the coverage of the RFID system is maximized and the accuracy of asset location identification is improved, especially for large warehouses or complex production environments. It can not only reduce scanning blind spots, but also reduce the workload of staff when identifying assets, reduce the number of times they need to approach assets or equipment, and improve work efficiency. At the same time, in order to address the problem of inaccurate asset inventory data caused by RFID identification penetrating space and identifying assets in other spaces, the present invention solves this problem through a signal strength RSSI (dBm) loss model. Usually, when such assets are scanned, the signal strength will be abnormally weakened. By checking whether the signal strength is abnormally weakened, it is determined whether it is an asset in other spaces and such assets are eliminated, thereby improving the accuracy of asset inventory. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0062] Figure 1 The present invention is a flowchart of a method for improving RFID recognition accuracy based on equidistant center calculation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0064] According to an embodiment of the present invention, a method for improving RFID recognition accuracy based on equidistant center calculation is provided.

[0065] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the method for improving RFID recognition accuracy based on equidistant center calculation according to an embodiment of the present invention, the method for improving RFID recognition accuracy based on equidistant center calculation includes the following steps:

[0066] S1. Obtain the equidistant center C and the average distance dis of the space.

[0067] In one embodiment, obtaining the equidistant centers and the average distances of the space includes the following steps:

[0068] S11. Select boundary points P: Select several points on the boundary of the space:

[0069] P 1 (x1,y1,z1),P 2 (x2,y2,z),...,P i (xi,yi,zi);

[0070] The equidistant centers are: C(cx,cy,cz).

[0071] S12. Construct the objective function and calculate the equidistant center of the space.

[0072] S13. Calculate the average distance in space using the Euclidean distance formula.

[0073] In one embodiment, constructing the objective function and calculating the equidistant center of the space includes the following steps:

[0074] S121, define the objective function;

[0075] Among them, the objective function is:

[0076] F(C)=Σ(d(C,P i ));

[0077] In the formula, F(C) represents the objective function, that is, the sum of the distances from each point to the center point C. By minimizing F(C), we can find the center point with the smallest total distance from the center point C to the boundary points of the space, that is, the equidistant center.

[0078] d(C,P i ) indicates the distance between the center point C and each point P i distance;

[0079] C refers to a specific center point;

[0080] P i are the boundary points selected above.

[0081] S122, using the gradient descent method to minimize the objective function and obtain the equidistant center. In order to calculate the equidistant center C using the gradient descent method, it is necessary to define a suitable loss function and then calculate its gradient with respect to the equidistant center C. Next, the equidistant center C is iteratively updated until the minimum value of the loss function is reached.

[0082] In one embodiment, minimizing the objective function using the gradient descent method and obtaining the isodistance center includes the following steps:

[0083] S1221, initialize d to a random value;

[0084] S1222. Define loss function

[0085] In the formula, d(C, P i ) indicates the distance between the center point C and each point P i The distance, n represents the number of data points, for example, P 1 , P 2 , ..., P n ;

[0086] S1223, calculating the gradient of the loss function with respect to d;

[0087] S1224, update d using the gradient descent method; that is, find the rate of change of the loss function with respect to the distance parameter d. In this case, it is necessary to calculate the objective function F(C) = ∑(d(C,P i ))The derivative or partial derivative with respect to d.

[0088] The calculation formula for updating is:

[0089]

[0090] In the formula, d represents the distance between the center point C and each point P i The distance between the two, a represents the learning rate. The learning rate is an important hyperparameter in the gradient descent algorithm, which is used to control the step size of parameter updates in each iteration. It determines the magnitude of each update in the gradient descent process. During the calculation process, it is necessary to conduct multiple judgments based on the actual situation and changes to set a reasonable learning rate;

[0091] L(d) represents the loss function, Indicates partial derivative operation on the variables in the function;

[0092] S1225, repeat S1223-S1224 until the preset number of iterations. Specifically, until it is repeated 10 times in a row, it is found that the gradient calculated in each S1223 does not change much, that is, the minimum value of the loss function is reached, then it is considered that the convergence condition is met. C in the last repetition is the equidistant center.

[0093] In one embodiment, calculating the gradient of the loss function with respect to d includes the following steps:

[0094] Expand the loss function And the distance from the center point C to each point P i Distance To find the derivative, T represents the transpose operation. Specifically, for a vector or matrix, it means converting a row vector into a column vector, or converting a column vector into a row vector.

[0095] Sum the obtained derivatives to get the gradient of the loss function L(d) with respect to d.

[0096] S2. Perform preliminary calibration on the RFID radio frequency power based on the average distance in space.

[0097] In one embodiment, the preliminary calibration of the RFID radio frequency power based on the average distance in space includes the following steps:

[0098] S21, setting the average distance in the space as the (optimal) identification distance of RFID identification, thereby calculating the optimal RFID radio frequency power. Identification distance (b) = average distance (dis), and obtaining the radio frequency power through the radio frequency power calculation formula;

[0099] The RF power calculation formula is:

[0100]

[0101] The theoretical basis of the above formula is based on the free space propagation model, in which the signal propagates in space without obstacles and interference. This model assumes that the attenuation of the RF signal in free space conforms to the inverse square law of the distance, and does not consider the influence of other factors, such as multipath effect, the influence of obstacles, etc. Therefore, this formula provides an approximate theoretical estimate that can help understand the maximum recognition distance of RFID under ideal conditions.

[0102] The RF power x (dBm) is calculated by the above formula; the free space propagation loss factor is taken as 20-30dB, except in extreme cases.

[0103] S22, adjusting the radio frequency power of the RFID identification device to the calculated radio frequency power.

[0104] S3. Adjust the RFID radio frequency power step by step.

[0105] In one embodiment, the step-by-step adjustment of the RFID radio frequency power includes the following steps:

[0106] S31. Assign the equidistant center of the space to the adjustment distance z(m): z(m)=d(m)

[0107] S32, placing the physical asset with the tag at the adjustment distance, and using the set RFID identification device to identify it;

[0108] S33. If it is not recognized, the RF power is increased by 0.1 and recognized again. If it is recognized, the RF power is reduced by 0.1 and recognized again. If it is recognized this time and not recognized in the next operation, the adjustment is completed.

[0109] S4. Calculate the signal strength RSSI (dBm) of RFID recognition using the loss model, and determine whether assets in other spaces are recognized in the current space.

[0110] The formula for the loss model is:

[0111] RSSI(dBm)=x(dBm)-20*log 10 (y)-20*log 10(f)+147.55;

[0112] Where RSSI (dBm) is the received signal strength.

[0113] x(dBm) is the RF power.

[0114] y is the transmission distance in meters.

[0115] f is the operating frequency in Hertz.

[0116] In one embodiment, receiving the signal strength of RFID identification and determining whether assets in other spaces are identified in the current space using a loss model includes the following steps:

[0117] S41, setting a number of calibrated signal strength values ​​for different transmission distances, that is, the calibrated RSSI value when the distance is d1 is RSSI1, the calibrated RSSI value when the distance is d2 is RSSI2, and so on, the calibrated RSSI value when the distance is dn is RSSIn;

[0118] Collect sample data corresponding to the transmission distance, that is, collect n sample data with distances of d1, d2, ..., dn, and substitute the transmission distance into the loss model to calculate the signal strength x and the average signal strength corresponding to each sample data. The standard deviation is calculated to obtain the standard deviation of the sample data; that is, the sample data is the signal strength calculated by substituting the transmission distance of each sample into the loss model, and the standard deviation is based on the standard deviation of these signal strengths.

[0119] S42: Obtain a threshold value to determine whether assets in other spaces are scanned.

[0120] In one embodiment, the threshold (ΔRSSI_threshold) is twice the standard deviation of the sample data ΔRSSI_threshold=2σ.

[0121] In one embodiment, when determining whether assets in other spaces are scanned, if the difference between the actual signal strength value and the calibrated signal strength value is greater than a threshold, it is (roughly) determined that the current space recognizes assets in other spaces.

[0122] The formula is:

[0123] ΔRSSI = actual RSSI - calibrated RSSI. If |ΔRSSI|>ΔRSSI_threshold, it may indicate that assets in other spaces have been scanned.

[0124] dBm is the RF power;

[0125] Free space propagation loss factor (usually 20-30 dB);

[0126] b is the recognition distance;

[0127] C is the equidistant center;

[0128] RSSI is the signal strength.

[0129] In summary, the method for improving RFID recognition accuracy based on equidistant center calculation of the present invention can help enterprises find the most suitable radio frequency strength for the current space before asset inventory, and solve the problem of cross-space scanning assets based on the current space and radio frequency strength. The present invention uses the equidistant center of the space as the farthest distance for RFID identification, which can ensure that the coverage of the RFID system is maximized and the accuracy of asset location identification is improved, especially for large warehouses or complex production environments. It can not only reduce the scanning blind area, but also reduce the work intensity of staff when identifying assets, reduce the number of times they need to approach assets or equipment, and improve work efficiency. At the same time, in order to solve the problem of inaccuracy of asset inventory data caused by RFID identification penetrating space and identifying assets in other spaces, the present invention solves it through the signal strength RSSI (dBm) loss model. Usually, when such assets are scanned, the signal strength will be abnormally weakened. By judging whether the signal strength is abnormally weakened, it is judged whether it is an asset in other spaces and such assets are eliminated, thereby improving the accuracy of asset inventory.

[0130] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for improving RFID recognition accuracy based on equidistant center calculation, characterized in that: The method for improving RFID recognition accuracy based on equidistant center calculation comprises the following steps: S1. Obtain the equidistant center and average distance of the space; S2. Preliminary adjustment of the RFID radio frequency power based on the average distance in space; S3, adjusting the radio frequency power of the RFID step by step; S4. Calculate the signal strength of RFID recognition using the loss model, and determine whether assets in other spaces are recognized in the current space; The acquisition of the equidistant centers and the average distance of the space comprises the following steps: S11, selecting a number of points on the boundary of the space; S12, construct the objective function and calculate the equidistant center of the space; S13, calculating the average distance in space by using the Euclidean distance formula; The construction of the objective function and calculation of the equidistant center of the space include the following steps: S121, define the objective function; Among them, the objective function is: F(C)=∑(d(C,P i )) Where F(C) represents the objective function; d(C,P i ) indicates the distance between the center point C and each point P i distance; S122. Use the gradient descent method to minimize the objective function and obtain the equidistant centers.

2. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 1, characterized in that: The method of minimizing the objective function using the gradient descent method and obtaining the equidistant centers includes the following steps: S1221, initialize d to a random value; S1222. Define a loss function; S1223, calculating the gradient of the loss function with respect to d; S1224, update d using the gradient descent method; The calculation formula for updating is: In the formula, d represents the distance between the center point C and each point P i distance; a represents the learning rate; L(d) represents the loss function, represents partial derivative operation; S1225. Repeat S1223-S1224 until the preset number of iterations is reached.

3. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 2 is characterized in that: The calculation of the gradient of the loss function with respect to d comprises the following steps: Expand the loss function and calculate the distance between the center point C and each point P i The distance is used to find the derivative; Sum the obtained derivatives to get the gradient of the loss function with respect to d.

4. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 3 is characterized in that: The preliminary calibration of the RFID radio frequency power based on the average distance in space includes the following steps: S21, setting the average distance of the space as the identification distance of RFID identification, and obtaining the radio frequency power through the radio frequency power calculation formula; The RF power calculation formula is: S22, adjusting the radio frequency power of the RFID identification device to the calculated radio frequency power.

5. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 4, characterized in that: The step-by-step adjustment of the radio frequency power of the RFID comprises the following steps: S31, assigning the equidistant center of the space to the adjustment distance; S32, placing the physical asset with the tag at the adjustment distance, and using the set RFID identification device to identify it; S33. If it is not recognized, the RF power is increased by 0.1 and recognized again. If it is recognized, the RF power is reduced by 0.1 and recognized again. If it is recognized this time and not recognized in the next operation, the adjustment is completed.

6. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 1, characterized in that: The method of calculating the signal strength of RFID recognition by using the loss model and determining whether assets in other spaces are recognized in the current space includes the following steps: S41, setting a number of calibration signal strength values ​​for different transmission distances, and collecting sample data corresponding to the transmission distances, averaging and standard deviation of all sample data, and obtaining the standard deviation of the sample data; S42, obtaining a threshold value to determine whether assets in other spaces are scanned; Among them, the formula of the loss model is: RSSI(dBm)=x(dBm)-20*log10(y)-20*log 10 (f)+147.55; Where RSSI (dBm) is the received signal strength; x (dBm) is the RF power; y is the transmission distance; f is the operating frequency.

7. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 6, characterized in that: The threshold is twice the standard deviation of the sample data.

8. The method for improving RFID recognition accuracy based on equidistant center calculation according to claim 6, characterized in that: When determining whether assets in other spaces are scanned, if the difference between the actual signal strength value and the calibrated signal strength value is greater than a threshold, it is determined that the current space recognizes assets in other spaces.

Images