TWI887669B - Asset location detection method and radio frequency tag location determination system using the detection method - Google Patents

Abstract

The present invention provides an asset location detection method and a radio frequency tag location determination system applicable to the detection method. The asset location detection method comprises: using a reader to read multiple radio frequency tags of multiple assets set in a target cabinet at maximum power to obtain a current sensing data set, obtaining a historical sensing data set of the target cabinet from a cloud database, dividing the current sensing data set into a first training data subset (First Translation Set, FTS) and a first validation data subset (a First Validation Set, FVS) at a specific ratio based on an automated machine learning algorithm, and training a first location detection model (First Location Model, FLM) according to the first training data subset (FTS). Furthermore, the RF tag location determination system includes a plurality of RF tags, an RF tag reading module, a local computing and storage device, a cloud server, and a cloud database. The RF tag reading module includes a reader, at least one antenna electrically connected to the reader, and a transmission power regulator. The reader receives the reading data transmitted by the plurality of RF tags.

Description

Asset location detection method and radio frequency tag location determination system using the detection method

The present invention relates to the technical field of radio frequency tag reading and identification, and in particular to an asset location detection method and a radio frequency tag location determination system applicable to the detection method.

In recent years, the demand for big data and cloud computing services has grown rapidly, and data centers have expanded in size and equipped with more IT equipment to meet such demand. Therefore, the number of assets has increased with the increase in equipment, and an effective equipment inventory system is needed to manage equipment assets. Traditional barcodes or two-dimensional codes (QR codes) are only useful within the line of sight and have no memory to record equipment-related information. Therefore, it is gradually replaced by radio frequency tag (RFID) technology, which uses radio waves to exchange information between readers and tags.

Equipment asset tracking can be divided into two types: mobile asset tracking and static asset tracking. Mobile asset tracking uses fixed readers, usually installed indoors, especially in places such as import and export gantries to track mobile assets. Static asset tracking is that operators bring handheld readers to the cabinet and read the static assets in the cabinet. Although the handheld method is time-consuming and labor-intensive, it is a very common method today.

Theoretically, RF tag technology has the possibility of misreading or missing, which will affect the accuracy of RF tags and lead to inaccurate equipment inventory. Misreading is the reader reading a tag outside the area. It is an unexpected result of the tag sensitivity being too high or the radio wave being reflected, diffracted or scattered in a metal environment. The common method to improve misreading is to analyze the RF tag read count (Tag Read Count, TRC) and received signal strength indicator (Received Signal Strength Indicator, RSSI). The read count (TRC) information shows the number of times the tag is correctly read, and the received signal strength indicator (RSSI) is related to the distance between the tag and the antenna.

On the other hand, missed reads are when the tag in the area is not read by the reader. They are caused by destructive radio wave interference, tag antenna misalignment, poor tag antenna pattern, low tag sensitivity or other factors. Improving missed reads can be achieved by increasing the power of the reader, improving the tag sensitivity, or moving the tag from the blind area to the non-blind area.

However, to date, various conventional methods for detecting assets or objects stored in metal cabinets are still subject to the possibility of misreading or missing reading from time to time, and the detection accuracy is low and still cannot reach a satisfactory level. In particular, in recent years, with the increasing demand for cloud services, inventory management systems that can effectively and accurately detect, for example, information technology (IT) assets have become increasingly important.

Therefore, the industry is looking forward to developing an asset location detection method and an RF tag location determination system that can accurately determine the location of RF tags, solve the inconvenience of previous technologies that require manual handheld readers to read one by one, gradually reduce the probability of misreading and missed reading, shorten the reading time, and increase the accuracy of inventory equipment.

In view of the above-mentioned defects or problems of the prior art, the inventors of the present invention have therefore reviewed various feasible solutions. After repeated improvements, corrections and research and development, they have finally completed an asset location detection method and a radio frequency tag location determination system applicable to the detection method of the present invention; it does not require manual handheld readers to read one by one, which can not only solve the inconvenience defects of the prior art, but also shorten the reading time, gradually reduce the probability of misreading and missing reading, and has excellent characteristics such as being able to accurately determine the location of radio frequency tags and increase the accuracy of inventory equipment.

That is, according to the technical idea of one viewpoint of the present invention, a method for detecting the location of an asset can be provided, comprising: a step of reading a current sensing data set, using a reader to read a plurality of RF tags of a plurality of assets disposed in a target cabinet at maximum power to obtain a current sensing data set, wherein the current read data set at least includes: the location of the target cabinet, maximum power, tag read count (Tag Read Count, TRC), and tag received signal strength (Received Signal Strength The method comprises the steps of: receiving a historical sensing data set, obtaining a historical sensing data set of the target cabinet from a cloud database, wherein the historical sensing data set at least includes: information about the target cabinet location, maximum power, tag read count (TRC), tag RSSI, tag type, and tag location; and training a location detection model, based on an automated machine learning algorithm, dividing the current sensing data set into a first training data subset (First Translation Set, FTS) and a first validation data subset (a First Validation Set, FVS) in a specific ratio, and training a first location detection model (First Location Detection Model) according to the first training data subset (FTS). Model, FLM), wherein the automated machine learning algorithm includes KNN, SVM, and/or DNN; a location detection model verification step, based on accuracy, precision, recall, and F1-Score, using a confusion matrix to verify the prediction ability of the first location detection model; and based on the first location detection model (FLM), the first verification data subset (FVM), based on the read count and/or (RSSI), determine whether an abnormal event occurs to generate a first judgment result and output the first judgment result.

Furthermore, in the asset location detection method of the present invention, it is preferred that the method further includes: Record data update step: recheck the status of the assets and update the record data of the assets stored in a cloud database; database information synchronization step: synchronize the record data of the assets stored in the local database with the record data of the assets stored in the cloud database.

Furthermore, according to an embodiment of the present invention, in the asset location detection method of the present invention, it is preferred to further include: a radio frequency tag information establishment step: establishing a radio frequency tag with the information of the asset, and setting the radio frequency tag on the asset.

Furthermore, according to one embodiment of the present invention, in the asset location detection method of the present invention, it is preferred that the reading device adjustment step is to increase or decrease the power of the electromagnetic wave.

Furthermore, according to one embodiment of the present invention, in the asset location detection method of the present invention, it is preferred that the comparison in the read data comparison step is consistent, including whether there is misreading and/or missing reading.

Furthermore, according to one embodiment of the present invention, in the asset location detection method of the present invention, it is preferred that the reading device adjustment step is to reduce the power of the electromagnetic wave when it is determined that there is an erroneous reading; and to increase the power of the electromagnetic wave when it is determined that there is no erroneous reading but there is a missed reading.

On the other hand, that is, according to the technical idea of another viewpoint of the present invention, a radio frequency tag location determination system applicable to the asset location detection method can also be provided, which includes: a plurality of radio frequency tags, which are set on a plurality of assets; an radio frequency tag reading module, which includes a reader, a transmitting source, a plurality of antennas corresponding to the plurality of assets and electrically connected to the reader, and a transmitting power adjuster, wherein In the invention, the plurality of antennas are arranged in a predetermined configuration to form an antenna array and emit electromagnetic waves to the radio frequency tags, and receive the read data transmitted by the radio frequency tags and transmit it to the reader; a local computing and storage device is electrically connected to the reader, and the local computing and storage device includes a processing unit and a local database electrically connected to the processing unit. The read data is transmitted from the reader to the local database. to the local computing and storage device, the local database storing the record data of the assets; a cloud server connected to the local computing and storage device via a network; and a cloud database connected to the local computing and storage device and the cloud server via the network, the cloud database storing the record data of the assets, the record data of the local database being synchronized with the cloud database The processing unit compares the read data with the recorded data in the local database, adjusts the emission power of the electromagnetic wave of the at least one antenna according to the comparison result, and transmits the comparison result to the cloud server; wherein the recorded data includes the identification information of the asset, the reading number information of the radio frequency tag, and the intensity information of the electromagnetic wave received by the radio frequency tag.

Furthermore, according to one embodiment of the present invention, in the radio frequency tag position determination system of the present invention, it is preferred that the antennas emit electromagnetic waves in sequence one at a time.

Furthermore, according to one embodiment of the present invention, in the RF tag location determination system of the present invention, it is preferred that the antennas have equal distances.

Furthermore, according to one embodiment of the present invention, in the RF tag location determination system of the present invention, it is preferred that the height of the array of antennas is the same as the height of the asset or equipment.

Therefore, by using the asset location detection method of the present invention, or by using the RF tag location determination system of the present invention to which the detection method is applied, the read data read from the RF tag is compared with the record data in the local database to determine whether the location of the RF tag is consistent with the record data stored in the local database. According to the present invention, since the transmission power of the antenna of the RF tag reading module can be adjusted; and since the received read data can be compared with the record data stored in the local database, missed reading and misreading can be avoided.

In addition, since the number of RF tag readings and the received signal strength indication can be recorded and analyzed, even if there are still misreadings or missed readings after the antenna transmission power is adjusted within the rated range, it is only necessary to check whether the asset is lost or placed in the wrong location, and then update the recorded data in the cloud database and synchronize the recorded data in the local database.

Therefore, according to an asset location detection method of the present invention, or according to the RF tag location determination system of the present invention that applies the detection method, there is no need for manual handheld readers to read one by one, which can not only solve the inconvenience defects of the previous technology, but also shorten the reading time, gradually reduce the probability of misreading and missed reading, and have excellent characteristics such as being able to accurately determine the location of the RF tag and increase the accuracy of the inventory equipment.

10: RF tag

20: RF tag reading module

21: Reader

22: Antenna

23: Transmit power adjuster

24: Emission source

30: Local computing and storage devices

31: Processing unit

32: Local database

33: Display unit

34: Data analysis module

40:Gateway

50: Cloud Server

60: Cloud database

N: Network

A1~A7: Steps

S1~S9: Steps

FIG1 is a schematic diagram of an embodiment of the RF tag location determination system of the present invention, in which an RF tag reading module and a local computing and storage device are connected to read multiple RF tags.

FIG2 is a block diagram of the first embodiment of the radio frequency tag position determination system of the present invention.

FIG3 is a block diagram of the second embodiment of the radio frequency tag position determination system of the present invention.

FIG4 is a block diagram of the third embodiment of the radio frequency tag position determination system of the present invention.

FIG5 is a block diagram of the fourth embodiment of the radio frequency tag position determination system of the present invention.

FIG6 is a block diagram of the fifth embodiment of the radio frequency tag position determination system of the present invention.

FIG7 is a block diagram of the sixth embodiment of the radio frequency tag position determination system of the present invention.

Figure 8 is a flowchart of machine training for data analysis using local computing and storage devices.

Figure 9 is a performance comparison chart of multiple machine learning models and feature data obtained from reading information.

Figure 10 is a flow chart of an embodiment of the asset location detection method created by the present invention.

Please refer to Figure 1. Figure 1 shows a specific example of the RF tag location determination system of the present invention. As shown in Figure 1, in this specific example, at least one RF tag 10 is set on an asset storage cabinet, and a reader 21, at least one antenna 22 electrically connected to the reader 21, and a local computing and storage device 30 electrically connected to the reader 21 are set on a machine. In this specific example, there is no special restriction on the number of RF tags 10, for example, 2 or more can be set. The asset storage cabinet can be, for example, equipment set in different places and locations; there is no special restriction on the number of the asset storage cabinets, for example, it can be at least one or more.

In this specific example, the at least one antenna 22 transmits electromagnetic waves to the at least one RF tag 10, and the RF tag 10 is driven by the electromagnetic waves; then, the RF tag 10 transmits the stored data to the reader 21, and the reader 21 receives and reads the stored data stored in multiple RF tags in multiple asset storage cabinets located at different locations to form multiple read data, and the read data is transmitted to a local computing and storage device 30. The RF tag reading module 20, the reader and the local computing and storage device 30 form a communication connection, and the connection means are not limited, for example, they can be connected by wired or wireless transmission.

Furthermore, in this specific example, the RF tag 10 can be a passive RF tag, that is, it generates electricity after receiving external electromagnetic waves and transmits the stored data. Each RF tag 10 in this specific example has multiple stored data, such as information about itself and the assets, such as the location or installation location of each RF tag 10; information such as the type, model, size, and item name of each asset or equipment.

Next, please refer to FIG. 2 and FIG. 3. FIG. 2 and FIG. 3 are block diagrams showing the RF tag location determination system of the first embodiment and the second embodiment of the present invention, respectively. In the first embodiment and the second embodiment, the RF tag location determination system of the present invention comprises at least a plurality of RF tags 10, an RF tag reading module 20, a local computing and storage device 30, a gateway 40, a cloud server 50 and a cloud database 60; wherein the RF tag 10, the RF tag reading module 20 and the local computing and storage device 30 are electrically connected, for example, by wired or wireless transmission. The local computing and storage device 30 is connected to a network N via a gateway 40, and is connected to a cloud server 50 and a cloud database 60 via the network N.

Multiple RF tags 10 are set on multiple assets at different locations, such as multiple different devices. Each RF tag 10 has multiple stored data, such as relevant information of the assets it is set on, such as the type, model, machine location or setting location information of each device. The RF tag 10 of this embodiment is a passive RF tag, that is, it generates electricity after receiving external electromagnetic waves and transmits the stored data.

In the first embodiment and the second embodiment, the RF tag reading module 20 includes a reader 21, at least one antenna 22 electrically connected to the reader 21, a transmission power adjuster 23, and a transmission source 24. According to the present invention, at least one antenna 22 transmits electromagnetic waves to a plurality of RF tags 10, and the plurality of RF tags 10 are driven by the electromagnetic waves to transmit their stored data to the reader 21, and receive the read data transmitted by the plurality of RF tags 10 and transmit it to the reader 21. The transmitting source 24 is connected to the antenna 22, and the transmitting power adjuster 23 is connected to the transmitting source 24. The transmitting source 24 is fed with an alternating voltage, which forms an electromagnetic wave through the antenna 22 and is transmitted to the RF tag 10. The transmitting power adjuster 23 can adjust the voltage value or alternating frequency fed from the transmitting source 24 to the antenna 22, thereby changing the power of the electromagnetic wave.

In the first embodiment and the second embodiment, the local computing and storage device 30 is electrically connected to the reader 21, and the local computing and storage device 30 includes a processing unit 31 and a local database 32 electrically connected to the processing unit 31. The local database 32 stores the record data of multiple assets. Multiple read data are transmitted from the reader 21 to the local computing and storage device 30. The processing unit 31 compares the read data with the record data in the local database 32 to determine whether the RF tag 10 is missed or misread.

In addition, in the second embodiment, in addition to all the components of the first embodiment, the RF tag location determination system of the present invention further includes a gateway 40. Therefore, in the second embodiment, the local computing and storage device 30 can be further connected to a network N via the gateway 40, and connected to the cloud server 50 and the cloud database 60 via the network N.

The third and fourth embodiments of the present invention are described below with reference to FIG. 4 and FIG. 5. As shown in FIG. 4 and FIG. 5, block diagrams of the RF tag position determination system of the third and fourth embodiments of the present invention are shown respectively.

In the third and fourth embodiments, the RF tag location determination system of the present invention comprises at least a plurality of RF tags 10, an RF tag reading module 20, a local computing and storage device 30, a cloud server 50, and a cloud database 60; wherein the RF tag reading module 20 further comprises a reader 21, at least one antenna 22 electrically connected to the reader 21, a transmission power adjuster 23, and a transmission source 24; and the local computing and storage device 30 further comprises a processing unit 31, and a local database 32 electrically connected to the processing unit 31.

In the third and fourth embodiments, the RF tag 10, the RF tag reading module 20 and the local computing and storage device 30 are electrically connected; the local computing and storage device 30 is electrically connected to the reader 21, is communicatively connected to a network N, and is connected to the cloud server 50 and the cloud database 60 via the network N. In addition, the transmitting source 24 is connected to the antenna 22, and the transmitting power adjuster 23 is connected to the transmitting source 24.

In the third and fourth embodiments, at least one antenna 22 transmits electromagnetic waves to a plurality of RF tags 10, and the plurality of RF tags 10 are driven by the electromagnetic waves to transmit their stored data to the reader 21, and receive the read data transmitted by the plurality of RF tags 10 and transmit it to the reader 21. The transmitting source 24 feeds an alternating voltage, which is formed into an electromagnetic wave through the antenna 22 and transmitted to the RF tag 10. The transmitting power adjuster 23 can adjust the voltage value or alternating frequency fed by the transmitting source 24 to the antenna 22, thereby changing the power of the electromagnetic wave. In addition, the local database 32 stores the record data of a plurality of assets. A plurality of read data are transmitted from the reader 21 to the local computing and storage device 30. The processing unit 31 compares the read data with the recorded data in the local database 32 to determine whether the RF tag 10 is missed or misread.

Furthermore, in the fourth embodiment, in addition to having all the components of the first embodiment, the local computing and storage device 30 in the RF tag position determination system of the present invention further includes a display unit 33. Therefore, in the fourth embodiment, in addition to being transmitted from the reader 21 to the local computing and storage device 30 as in the first embodiment, the plurality of read data can also be further transmitted to the display unit 33. In addition to comparing the read data with the record data in the local database 32 as in the first embodiment to determine whether there is a missed or misread RF tag 10, the processing unit 31 can also further display it on the display unit 33.

In addition, in the fourth embodiment, in addition to all the components of the third embodiment, the radio frequency tag location determination system of the present invention further includes a gateway 40. Therefore, in the fourth embodiment, the local computing and storage device 30 can be further connected to a network N via the gateway 40, and connected to the cloud server 50 and the cloud database 60 via the network N.

Next, the fifth and sixth embodiments of the present invention will be described with reference to FIG6 and FIG7. As shown in FIG6 and FIG7, they are block diagrams showing the RF tag position determination system of the fifth and sixth embodiments of the present invention, respectively.

In the fifth and sixth embodiments, the RF tag location determination system of the present invention comprises at least a plurality of RF tags 10, an RF tag reading module 20, a local computing and storage device 30, a cloud server 50, and a cloud database 60, respectively; wherein, the RF tag reading module 20 comprises a reader 21, at least one antenna 22 electrically connected to the reader 21, a transmission power adjuster 23, and a transmission source 24; the local computing and storage device 30 comprises a processing unit 31, a local database 32 electrically connected to the processing unit 31, a display unit 33, and a data analysis module 34, and the data analysis module 34 can be a program module. In addition, in the sixth embodiment, the RF tag position determination system of the present invention further includes a gateway 40.

Furthermore, in the fifth and sixth embodiments, the RF tag 10, the RF tag reading module 20 are electrically connected to the local computing and storage device 30; the local computing and storage device 30 is connected to a network N via a gateway 40, and is connected to a cloud server 50 and a cloud database 60 via the network N. The transmitting source 24 is connected to the antenna 22, and the transmitting power adjuster 23 is connected to the transmitting source 24. The local computing and storage device 30 is electrically connected to the reader 21.

Furthermore, in the fifth and sixth embodiments, the local database 32 stores the record data of multiple assets. At least one antenna 22 transmits electromagnetic waves to multiple RF tags 10, and the multiple RF tags 10 are driven by the electromagnetic waves to transmit their stored data to the reader 21, and receive the read data transmitted by the multiple RF tags 10 and transmit them to the reader 21. The transmitting source 24 is fed with alternating voltage, which is formed into electromagnetic waves through the antenna 22 and transmitted to the RF tags 10. The transmitting power adjuster 23 can adjust the voltage value or alternating frequency fed by the transmitting source 24 to the antenna 22, thereby changing the power of the electromagnetic waves. A plurality of read data are transmitted from the reader 21 to the local computing and storage device 30 and displayed on the display unit 33. The processing unit 31 compares the read data with the record data in the local database 32 to determine whether there is a missed or misread RF tag 10, and the comparison result is also displayed on the display unit 33.

In addition, in the fifth and sixth embodiments, the data analysis module 34 in the local computing and storage device 30 of the present invention can be a program module. Moreover, according to the present invention, the data analysis module 34 of the local computing and storage device 30 can use various learning models for learning, so that the data analysis module 34 can accurately analyze the received read data. Moreover, there is no special limitation on the models that can be applied to the data analysis module 34 of the present invention, for example, Support Vector Machine (SVM), K-Nearest Neighbor and Deep Neural Network (DNN), etc.

Furthermore, in the sixth embodiment, in addition to all the components of the fifth embodiment, the radio frequency tag location determination system of the present invention further includes a gateway 40. Therefore, in the sixth embodiment, the local computing and storage device 30 can be further connected to a network N via the gateway 40, and connected to the cloud server 50 and the cloud database 60 via the network N.

The following further describes the terminology definitions, related actions, operation methods, and related variations of the RF tag position determination system of the first, second, third, fourth, fifth, and sixth embodiments of the present invention.

The "read data" referred to in the present invention includes the identification information of the asset, the number of read times of the RF tag 10, and the intensity information of the electromagnetic wave received by the RF tag 10.

The "asset identification information" referred to in the present invention is the same as the stored data of the RF tag 10, which can be the type, model, machine location or setting location information of the equipment. The reading number information of the RF tag 10 and the intensity information of the electromagnetic wave received by the RF tag 10 are established after the reader 21 reads.

The "tag read count information (TRC) of the RF tag 10" referred to in the present invention is the sum of the number of times the RF tag is read by the reader. Multiple RF tags can respond to the reader at the same time in space, but only the RF tag that receives the reader's response can perform subsequent communications. The RF tag's received electromagnetic wave strength information (Received Signal Strength Indicator, RSSI) measures the power of the signal received by the reader. The RF tag's received electromagnetic wave strength information indicates the distance between the RF tag and the reader.

The local computing and storage device 30 is electrically connected to the reader 21. The local computing and storage device 30 includes a processing unit 31, a local database 32 electrically connected to the processing unit 31, and a display unit 33. The local database 32 stores the record data of multiple assets. The "reading times information of the RF tag 10" and the "intensity information of the electromagnetic wave received by the RF tag 10" referred to in the present invention are respectively established after the reader 21 reads.

The "recorded data" referred to in the present invention may be, for example, the type, model, machine location or setting location information of the aforementioned equipment, as well as the number of times the RF tag 10 was read during the previous detection, and the intensity information of the electromagnetic wave received by the RF tag 10. A plurality of read data are transmitted from the reader 21 to the local computing and storage device 30 and displayed on the display unit 33. The processing unit 31 compares the read data with the recorded data in the local database 32 to determine whether there is a missed reading of the RF tag 10 or a misreading of the RF tag 10, and the comparison result is also displayed on the display unit 33.

The term "missed reading" in the present invention means that the recorded data in the local database 32 contains data of a certain RF tag, but the read data does not contain the data of the RF tag.

The "misreading" referred to in the present invention is that the reading frequency information of the RF tag 10 and the intensity information of the electromagnetic wave received by the RF tag 10 in the read data are inconsistent with the reading frequency information of the RF tag 10 or the intensity information of the electromagnetic wave received by the RF tag 10 in the recorded data, which means that the RF tag 10 and the reader 21 are located at different locations, and the RF tag 10 of another device should be read, or the device corresponding to the RF tag 10 has been moved.

Furthermore, according to the present invention, the multiple RF tags 10 in the first, second, third, fourth, fifth, and sixth embodiments may be multiple identical or different assets respectively installed in different places. For example, the assets may be various equipment such as asset lockers, storage boxes, warehouses, and warehouses. Each RF tag 10 has multiple stored data, such as information related to the assets it is installed in, such as the type, model, machine location, or installation location information of each equipment.

Furthermore, according to the present invention, the RF tag 10 can be a passive RF tag, that is, it generates electricity after receiving external electromagnetic waves and transmits the stored data.

Furthermore, according to the present invention, when the read data does not match the recorded data, the transmission power adjuster 23 of the RF tag reading module 20 is operated to adjust the power of the electromagnetic wave emitted by the antenna 22, for example, to increase or decrease the transmission power, and then the RF tag 10 is read again to generate the read data. Then the re-read read data is compared with the recorded data in the local database 32. If there is still a discrepancy, the power of the electromagnetic wave emitted by the antenna 22 is adjusted again, and the reading and comparison are continued until all possible electromagnetic wave powers within the adjustment range of the transmission power adjuster 23 have been implemented. If there is still a discrepancy between the read data and the recorded data, it means that the asset is lost or the moved location is not recorded.

In addition, the local computing and storage device 30 of the present invention further includes a data analysis module 34, which can be a program module. Various learning models are used to provide the data analysis module 34 of the local computing and storage device 30 for learning, so that the data analysis module 34 can accurately analyze the received read data. The models used in this embodiment include Support Vector Machine (SVM), K-Nearest Neighbor and Deep Neural Network (DNN).

In the first, second, third, fourth, fifth and sixth embodiments, the cloud server 50 is connected to the local computing and storage device 30 via a network N. The cloud database 60 is also connected to the local computing and storage device 30 and the cloud server 50 via the network N. The cloud database 60 also stores multiple records of multiple assets, and the multiple records of the local database 32 are synchronized with the multiple records of the cloud database 60. If the read data does not match the recorded data after continuously adjusting the electromagnetic wave transmission power of the antenna 22, the local computing and storage device 30 will transmit the comparison result and the read data to the cloud server 50 via the network N and issue a warning.

After the status of the assets or equipment on site is physically investigated by personnel, the record data of the cloud database 60 is updated according to the latest status of the assets or equipment, and the record data of the local database 32 is also updated synchronously through the connection between the local computing and storage device 30 and the cloud database 60 for the next inspection.

Furthermore, in the first, second, third, fourth, fifth, and sixth embodiments, the plurality of antennas 22 of the RF tag reading module 20 may be arranged in an array, and the configuration of the array of antennas 22 may correspond to the configuration of the asset or device, for example, the array of antennas 22 may be arranged perpendicular to the ground or parallel to the ground or arranged in an arc, and may be modified according to the configuration of the asset or device.

According to the present invention, the antennas 22 in the array of antennas 22 have equal distances from each other, and the height of the array of antennas 22 is equal to the height of the asset or equipment. The antennas 22 in the array of antennas 22 emit electromagnetic waves in sequence one at a time. In this way, multiple antennas 22 emit electromagnetic waves at different positions, which can prevent the RF tag 10 from not receiving electromagnetic waves and being unable to transmit stored data.

According to the present invention, in another embodiment, the antenna 22 can also be configured to be movable, and can move along the predetermined track of the bracket, and emit electromagnetic waves every time it moves to a certain point, which can also achieve the same effect as the antenna array.

The following is an explanation of the asset location detection method of the present invention. In one embodiment, the asset location detection method of the present invention may include: a step of reading a current sensing data set, using a reader to read multiple RF tags of multiple assets set in a target cabinet at maximum power to obtain a current sensing data set, wherein the current read data set at least includes: the target cabinet location, maximum power, tag read count (Tag Read Count, TRC), and tag received signal strength (Received Signal Strength The method comprises the steps of: receiving a historical sensing data set, obtaining a historical sensing data set of the target cabinet from a cloud database, wherein the historical sensing data set at least includes: information about the target cabinet location, maximum power, tag read count (TRC), tag RSSI, tag type, and tag location; and training a location detection model, based on an automated machine learning algorithm, dividing the current sensing data set into a first training data subset (First Translation Set, FTS) and a first validation data subset (a First Validation Set, FVS) in a specific ratio, and training a first location detection model (First Location Detection Model) according to the first training data subset (FTS). Model, FLM), wherein the automated machine learning algorithm includes KNN, SVM, and/or DNN; a location detection model verification step, based on accuracy, precision, recall, and F1-Score, using a confusion matrix to verify the prediction ability of the first location detection model; and based on the first location detection model (FLM), the first verification data subset (FVM), based on the read count and/or (RSSI), determine whether an abnormal event occurs to generate a first judgment result and output the first judgment result.

More specifically, an embodiment of the asset location detection method of the present invention is described with reference to FIG. 10. In this embodiment, the aforementioned steps of reading the current sensing data set and receiving the historical sensing data set can be further divided into: an RF tag information establishment step (hereinafter referred to as "step S1"), an RF tag reading step (hereinafter referred to as "step S2"), a read data comparison step (hereinafter referred to as "step S3"), and a comparison result generation step. Matching reporting step (hereinafter referred to as "step S4"), reading device adjustment step (hereinafter referred to as "step S5", "step S6"), comparison result reporting step (hereinafter referred to as "step S7"), record data update step (hereinafter referred to as "step S8"), database information synchronization step (hereinafter referred to as "step S9").

In step S1 (RF tag information establishment step), an RF tag is established with the information of an asset, and the RF tag is set on the asset. The asset information can be the type, model, machine location or setting location information of the aforementioned equipment, etc. The asset information is stored in the RF tag, and then the RF tag is attached to the corresponding asset, such as on the back of the equipment. Then proceed to step S2.

In step S2 (RF tag reading step), electromagnetic waves are emitted to the RF tag to drive the RF tag to transmit its stored data to the reader, so as to read the stored data stored in multiple RF tags of multiple assets set at different locations to form multiple read data. Then proceed to step S3.

In step S3 (reading data comparison step), multiple read data are compared with the record data of multiple assets pre-stored in a local database to see if they match. If they match, it means there is no misreading or missing reading, and then proceed to step S4. If they do not match, it is judged to be missing reading or misreading. The judgment method of missing reading or misreading is the same as the method of explaining the radio frequency tag position determination system of the present invention. If it is a missing reading, proceed to step S5, and if it is a misreading, proceed to step S6.

In step S4 (comparison result matching reporting step), the matching result between the read data and the record data in the local database is transmitted to a cloud server.

In step S5 (reading device adjustment step) or step S6 (reading device adjustment step), if the read data does not match the recorded data, the power of the electromagnetic wave emitted by the antenna for reading multiple radio frequency tags is adjusted. That is, according to the reading device adjustment step of the present invention, since the power of the electromagnetic wave can be adjusted, the occurrence of misreading or missed reading can be reduced.

For example, if it is a misread, it may be that the power of the electromagnetic wave emitted by the antenna is too high, so that the RF tag of other devices is read, so the power of the electromagnetic wave can be reduced. If it is a missed read, it may be that the power of the electromagnetic wave emitted by the antenna is too low, so that the RF tag of the current device cannot be read, so the power of the electromagnetic wave can be increased.

In step S7 (comparison result reporting step), if the read data and the recorded data do not match within the power adjustment range of the electromagnetic wave, the comparison result of the read data and the recorded data is transmitted to the cloud server via the network.

In step S8 (record data update step), the status of the asset is rechecked, and the record data of the asset stored in a cloud database is updated. That is, according to the record data update step of the present invention, the status of the assets can be rechecked, and the record data of the assets stored in a cloud database can be updated.

In step S9 (database information synchronization step), the record data of the assets stored in the local database is synchronized with the record data of the assets stored in the cloud database. That is, according to the database information synchronization step of the present invention, the record data of the assets stored in the local database can be synchronized with the record data of the assets stored in the cloud database.

The following describes the machine learning stage in the asset location detection method of the present invention. According to the technical idea of the present invention, the above machine learning stage can be divided into a location detection model training step and a location detection model verification step.

In one embodiment, the location detection model training step of the present invention is to divide the current sensing data set into a first training data subset (First Translation Set, FTS) and a first validation data subset (a First Validation Set, FVS) in a specific ratio based on an automated machine learning algorithm, and train a first location detection model (First Location Model, FLM) according to the first training data subset (FTS), wherein the automated machine learning algorithm includes KNN, SVM, and/or DNN; the location detection model verification step of the present invention verifies the prediction ability of the first location detection model based on accuracy, precision, recall, and F1-Score using a confusion matrix.

Furthermore, in one embodiment, the present invention can also determine whether an abnormal event occurs based on the first location detection model (FLM), the first verification data subset (FVM), and the read count and/or RSSI to generate a first judgment result and output the first judgment result.

For example, 200 IQ350 RF tags can be placed in a metal equipment cabinet in 40 rows and 5 columns as the object of the machine learning training of the present invention. For example, in the machine learning stage of the present invention, according to the machine training process shown in FIG8, the data of each RF tag is read by a Zebra Technology FX9600 reader in conjunction with a KAPC-08CF linear antenna. Then, the above three learning models are used to analyze the aforementioned tag read count information (TRC) and the RF tag received electromagnetic wave strength information (RSSI) to train the local computing and storage device 30.

Then, the results are recorded in Table 1 below and Figure 9 shows a comparison of the performance of three learning models, Support Vector Machine (SVM), K-Nearest Neighbor (K-Nearest Neighbor) and Deep Neural Network (DNN), relative to the features TRC and RSSI in the used data.

Furthermore, as shown in Table 1 and FIG. 9 , it can be confirmed that the combination of RSSI, TRC and SVM provides the most significant prediction performance with an accuracy of 99.4%. Other combinations, including RSSI, KNN and DNN, also perform well. Among the three models, SVM has the best performance. . Because RSSI provides more distance information between the RF tag and the reader antenna than TRC, RSSI provides more information about location classification than TRC, that is, according to the present invention, in one embodiment, the performance of using RSSI can obtain better performance than using TRC.

The following describes the machine training process of the machine learning stage in the asset location detection method of the present invention while referring to FIG8. In step A1, select the RF tag, reader and antenna; in step A2, optimize the reading distance and electromagnetic wave transmission power; in step A3, check whether the read data is correct. If erroneous data is found, return to step A1; if the data is found to be correct, enter step A4; in step A4, obtain the reading frequency information and electromagnetic wave intensity information from the read data, and Start machine learning; in step A5, check whether the machine's learning performance after training meets the requirements. If not, return to step A4 and train with other models. If it meets the requirements, proceed to step A6; in step A6, remove the error data and set the device with the frequency label attached to the correct position. The subsequent device inspection can be performed using the trained model.

In addition, although the present invention illustrates the execution process of a new static tracking method. Although the above invention illustrates a single type of tag, mixed types of tags can be used in the same cabinet with only slight modifications, so their situations are also applicable to the method of the present invention.

Furthermore, by repeating the steps shown in FIG. 10 N times, N tag types can be distinguished. Therefore, the method of the present invention can achieve more accurate and efficient superior effects than the traditional handheld method. On the other hand, the type information is preferably stored in the electronic product code (EPC) so that the reader can identify the tag and then use the appropriate power value and the trained model.

Therefore, by using the asset location detection method of the present invention, or by using the RF tag location determination system of the present invention to which the detection method is applied, the read data read from the RF tag is compared with the record data in the local database to determine whether the location of the RF tag is consistent with the record data stored in the local database. According to the present invention, since the transmission power of the antenna of the RF tag reading module can be adjusted; and since the received read data can be compared with the record data stored in the local database, missed reading and misreading can be avoided.

In addition, since the number of RF tag readings and the received signal strength indication can be recorded and analyzed, even if there are still misreadings or missed readings after the antenna transmission power is adjusted within the rated range, it is only necessary to check whether the asset is lost or placed in the wrong location, and then update the recorded data in the cloud database and synchronize the recorded data in the local database.

Therefore, according to an asset location detection method of the present invention, or according to the RF tag location determination system of the present invention that applies the detection method, there is no need for manual handheld readers to read one by one, which can not only solve the inconvenience defects of the previous technology, but also shorten the reading time, gradually reduce the probability of misreading and missed reading, and have excellent characteristics such as being able to accurately determine the location of the RF tag and increase the accuracy of the inventory equipment.

However, the above is only the preferred embodiment of the present invention, and it cannot be used to limit the scope of the implementation of the present invention. That is, all simple equivalent changes and modifications made according to the scope of the patent application and the content of the invention description of the present invention are still within the scope of the present invention. In addition, any embodiment or patent application of the present invention does not need to achieve all the purposes, advantages or features disclosed by the present invention. In addition, the abstract and title are only used to assist in searching for patent documents, and are not used to limit the scope of the rights of the present invention. In addition, the terms "first" and "second" mentioned in this specification or patent application are only used to name the element or distinguish different embodiments or scopes, and are not used to limit the upper or lower limit of the number of elements.

10: RF tag

21: Reader

22: Antenna

30: Local computing and storage devices

Claims (10)

A method for detecting an asset location includes: A step of reading a current sensing data set, using a reader to read multiple radio frequency tags of multiple assets set in a target cabinet at maximum power to obtain a current sensing data set, wherein the current read data set at least includes: information related to the target cabinet location, maximum power, tag read count (TRC), and tag received signal strength (RSSI); the target cabinet is a metal container or a metal storage cabinet; A step of receiving a historical sensing data set, obtaining a historical sensing data set of a target cabinet from a cloud database, wherein the historical sensing data set at least includes: information related to the target cabinet location, maximum power, tag read count (TRC), tag RSSI, tag type, and tag location; A step of training a location detection model, based on an automated machine learning algorithm, dividing the current sensing data set into a first training data subset (First Translation Set, FTS) and a first validation data subset (a First Validation Set, FVS) in a specific ratio, and training a first location detection model (First Location Model, FLM) according to the first training data subset (FTS), wherein the automated machine learning algorithm includes KNN, SVM, and/or DNN; A location detection model verification step, based on accuracy, precision, recall, and F1-Score, uses a confusion matrix to verify the prediction capability of the first location detection model; and Based on the first location detection model (FLM), the first verification data subset (FVM), based on the read count and/or (RSSI), it is determined whether an abnormal event occurs to generate a first judgment result and output the first judgment result. The asset location detection method as described in claim 1 further includes: Record data update step: recheck the status of the assets and update the record data of the assets stored in a cloud database; Database information synchronization step: synchronize the record data of the assets stored in the local database with the record data of the assets stored in the cloud database. The asset location detection method as described in claim 1 further includes: a radio frequency tag information establishment step: establishing a radio frequency tag with the information of the asset, and setting the radio frequency tag on the asset. An asset location detection method as described in claim 1, wherein the reading device adjustment step is to increase or decrease the power of the electromagnetic wave. In the asset location detection method as described in claim 1, whether the comparison in the read data comparison step is consistent includes whether the comparison has misreading and/or missing reading. As described in claim 5, the reading device adjustment step is to reduce the power of the electromagnetic wave when it is determined that there is an erroneous reading; and to increase the power of the electromagnetic wave when it is determined that there is no erroneous reading but a missed reading. A radio frequency tag location determination system, comprising: A plurality of radio frequency tags, arranged on a plurality of assets; A radio frequency tag reading module, comprising a reader, a transmitting source, a plurality of antennas corresponding to the plurality of assets and electrically connected to the reader, and a transmitting power adjuster, wherein the plurality of antennas are arranged in an antenna array in a predetermined configuration and transmit electromagnetic waves to the radio frequency tags, and receive the reading data transmitted by the radio frequency tags and transmit it to the reader; A local computing and storage device, electrically connected to the reader, the local computing and storage device includes a processing unit and a local database electrically connected to the processing unit, the read data is transmitted from the reader to the local computing and storage device, the local database stores the record data of the assets; A cloud server, connected to the local computing and storage device via a network; and A cloud database, connected to the local computing and storage device and the cloud server via the network, the cloud database stores the record data of the assets, the record data of the local database is synchronized with the record data of the cloud database; wherein The processing unit compares the read data with the recorded data in the local database, adjusts the emission power of the electromagnetic wave of the at least one antenna according to the comparison result, and transmits the comparison result to the cloud server; The recorded data includes the identification information of the asset, the reading number information of the radio frequency tag, and the intensity information of the electromagnetic wave received by the radio frequency tag. As described in claim 7, the radio frequency tag location determination system, wherein the antennas emit electromagnetic waves in sequence one at a time. A radio frequency tag location determination system as described in claim 7, wherein the antennas are equidistant from each other. A radio frequency tag location determination system as described in claim 7, wherein the height of the array of antennas is the same as the height of the asset or equipment.