A logistics warehousing full life cycle management method based on RFID radio frequency identification

Abstract

The present invention relates to the field of data management technology, and in particular to a logistics warehousing full life cycle management method based on RFID radio frequency identification, comprising: collecting inventory data of each type of goods at each recording moment; obtaining a selection degree according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the association between each sigma point vector at the target recording moment and the inventory data at different historical recording moments; obtaining a storage sigma point vector according to the selection degree of each sigma point vector; and performing data prediction management according to the storage sigma point vector. The present invention eliminates a large number of redundant sigma point vectors, improves the accuracy of data prediction results, improves the efficiency of data prediction, and improves the management efficiency of the full life cycle of logistics warehousing.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a logistics warehousing full life cycle management method based on RFID radio frequency identification. Background Art

[0002] RFID technology is a radio frequency technology that uses the principle of electromagnetic fields to achieve two-way communication between devices through wireless radio frequency. When using RFID technology to perform life analysis in logistics warehousing systems, it is necessary to predict the inventory of each type of item in the logistics warehousing system in order to manage the relevant data in the logistics warehousing system and complete data management for the entire life cycle of logistics warehousing.

[0003] When predicting the inventory of item types in a logistics warehousing system, existing methods usually use the UKF (Unscented Kalman Filter) algorithm to predict the inventory of each item type; however, due to the different external environmental noise conditions of items in different locations in the storage environment, the traditional UKF unscented Kalman filter algorithm has a large number of redundant sigma point vectors, which reduces the accuracy of data prediction and thus reduces the management efficiency of the entire life cycle of logistics warehousing. Summary of the invention

[0004] The present invention provides a logistics warehousing full life cycle management method based on RFID radio frequency identification to solve the existing problem that objects at different positions in a warehousing environment are subject to different external environmental noise conditions, resulting in a large number of redundant sigma point vectors in a traditional UKF unscented Kalman filter algorithm.

[0005] The present invention provides a logistics warehousing full life cycle management method using RFID radio frequency identification, which adopts the following technical solutions:

[0006] The following steps are involved:

[0007] Collect inventory data for each type of goods at each recording moment;

[0008] According to the inventory data of each type of goods at each recording moment, several sigma point vectors at each recording moment are obtained through the UKF unscented Kalman filter algorithm; any recording moment is recorded as the target recording moment, and each recording moment before the target recording moment is recorded as the historical comparison recording moment; according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, the selection degree of each sigma point vector is obtained; according to the selection degree of each sigma point vector, all sigma point vectors are divided to obtain several storage sigma point vectors at each recording moment;

[0009] Data prediction management is performed based on several stored sigma point vectors at each recording moment.

[0010] Preferably, the selection degree of each sigma point vector is obtained according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, including the specific method of:

[0011] For any historical comparison recording time and any commodity type, according to the difference in inventory data between the target recording time and the historical comparison recording time, the historical inventory limit change factor of the commodity type at the target recording time is obtained;

[0012] According to the overall change trend of the inventory data of the goods type at all historical comparison recording moments before the target recording moment, the historical inventory stability of the goods type at the target recording moment is obtained;

[0013] According to the historical inventory limit change factor of the goods type at the target recording time and the correlation of the historical inventory stability, a vector combination is constructed to obtain the goods comparison vector at the target recording time;

[0014] For any sigma point vector at the target recording time, the similarity between the cargo comparison vector and the sigma point vector at the target recording time is obtained;

[0015] The last historical comparison record time before the target record time is recorded as the historical neighbor comparison time. The selection degree of the sigma point vector is obtained according to the difference in inventory data changes between the target record time and the historical neighbor comparison time of all goods types, as well as the similarity of goods inventory between the goods comparison vector and the sigma point vector.

[0016] Preferably, the method of obtaining the historical inventory limit change factor of the commodity type at the target recording time according to the difference in inventory data between the target recording time and the historical comparison recording time includes the following specific methods:

[0017] The absolute value of the difference in inventory data between the target recording time and the historical comparison recording time of the goods type is recorded as the inventory change value between the target recording time and the historical comparison recording time; the inventory change values ​​between the target recording time and all historical comparison recording times are linearly normalized, and each normalized inventory change value is recorded as the inventory change factor; the maximum value of the inventory change factors between the target recording time and all historical comparison recording times is recorded as the historical inventory limit change factor of the goods type at the target recording time.

[0018] Preferably, the method of obtaining the historical inventory stability of the goods type at the target recording time according to the overall change trend of the inventory data of all historical comparison recording times before the target recording time includes the following specific methods:

[0019]

[0020] In the formula, α represents the historical inventory stability of the commodity type at the target recording time; N represents the number of all historical reference recording times at the target recording time; b n Indicates the inventory data of the goods type at the nth historical comparison record time; It represents the mean of the inventory data of the commodity type at all historical comparison record moments; σ represents the standard deviation of the inventory data of the commodity type at all historical comparison record moments.

[0021] Preferably, the combination construction of vectors is performed according to the correlation between the historical inventory limit change factor and the historical inventory stability of the goods type at the target recording time to obtain the goods comparison vector at the target recording time, including the specific method of:

[0022] According to the historical inventory limit change factor and historical inventory stability of the goods type at the target recording time, the moment goods trend vector at the target recording time is obtained; the vector after the moment goods trend vector is transposed is recorded as the moment goods comparison vector;

[0023] According to the inventory data of each type of goods at the target recording time, the state vector is constructed through the UKF unscented Kalman filter algorithm, and the state vector is recorded as the moment-by-moment goods state vector of the target recording time; according to the moment-by-moment goods state vector of the target recording time, the state transfer matrix of the target recording time is obtained through the UKF unscented Kalman filter algorithm; the vector obtained by multiplying the moment-by-moment goods comparison vector and the state transfer matrix is ​​recorded as the goods comparison vector of the target recording time.

[0024] Preferably, the method of obtaining the cargo trend vector at the target recording time according to the historical inventory limit change factor and the historical inventory stability of the cargo type at the target recording time includes the following specific methods:

[0025] The product of the historical inventory limit change factor and the historical inventory stability is recorded as the type data trend of the goods type at the target recording time; the type data trends of all goods types at the target recording time are obtained; according to the type data trends of all goods types at the target recording time, a state vector is constructed through the UKF unscented Kalman filter algorithm, and the state vector is recorded as the momentary goods trend vector at the target recording time.

[0026] Preferably, the similarity between the cargo comparison vector and the sigma point vector at the target recording time is obtained by the specific method of obtaining the cargo inventory similarity between the cargo comparison vector and the sigma point vector:

[0027] The cosine similarity between the cargo comparison vector and the sigma point vector is recorded as the cargo inventory similarity between the cargo comparison vector and the sigma point vector.

[0028] Preferably, the selection degree of the sigma point vector is obtained according to the change difference of the inventory data of all goods types between the target recording time and the historical neighbor comparison time, and the goods inventory similarity between the goods comparison vector and the sigma point vector, including the specific method of:

[0029]

[0030] In the formula, γ represents the selection degree of the sigma point vector; τ represents the similarity of the goods inventory between the goods comparison vector and the sigma point vector; M represents the number of all goods types; b1 m b2 represents the inventory data of the mth commodity type at the target recording time; m represents the inventory data of the mth commodity type at the historical nearest neighbor comparison time; α m It represents the historical inventory stability of the mth commodity type at the target recording time; || represents the absolute value.

[0031] Preferably, the method of dividing all sigma point vectors according to the selection degree of each sigma point vector to obtain a plurality of storage sigma point vectors at each recording moment includes the following specific methods:

[0032] For any recording moment, the sequence of all sigma point vectors at the recording moment arranged in descending order according to the selection degree is recorded as the sigma sequence to be selected. For any two adjacent sigma point vectors in the sigma sequence, the absolute value of the difference in the selection degree between the two sigma point vectors is recorded as the division degree of the two sigma point vectors; among the division degrees of all any two adjacent sigma point vectors in the sigma sequence to be selected, the two sigma point vectors corresponding to the smallest division degree are recorded as the sigma point vectors to be divided; among the two sigma point vectors to be divided, the second sigma point vector to be divided is recorded as the termination sigma point vector; each sigma point vector between the first sigma point vector and the termination sigma point vector in the sigma sequence to be selected is recorded as the storage sigma point vector at the recording moment.

[0033] Preferably, the data prediction management is performed according to a number of storage sigma point vectors at each recording moment, and the specific method includes:

[0034] The storage sigma point vector at each recording moment is used as the sigma point vector actually used at each recording moment. According to the inventory data of each type of goods at each recording moment and the sigma point vector actually used at each recording moment, the predicted value of each type of goods at each recording moment is obtained through the UKF unscented Kalman filter algorithm. The predicted values ​​of all types of goods at each recording moment are input into the AnalyticDB database of the central computer to complete the full life cycle management of logistics warehousing.

[0035] The beneficial effects of the technical solution of the present invention are as follows: by analyzing the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, an appropriate amount of sigma point vectors are adaptively selected to perform data prediction management; first, according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, the selection degree of each sigma point vector is obtained, and the selection degree is used to describe the obvious degree of the data change characteristics when the corresponding sigma point vector produces abnormal changes to the inventory data of each type of goods, so that s The relationship between the sigma point vector and the change of the overall environmental data is easier to observe; then according to the selection degree of each sigma point vector, several storage sigma point vectors are obtained after division, and the storage sigma point vector is used to reflect the sigma point vectors with actual greater management value, excluding a large number of redundant sigma point vectors, thereby improving the efficiency of data prediction; the present invention analyzes the changes in inventory data of each cargo at different recording times, and combines the correlation between the sigma point vector and the inventory data to adaptively select an appropriate amount of sigma point vectors for data prediction management; a large number of redundant sigma point vectors are excluded, thereby improving the accuracy of data prediction results, improving the efficiency of data prediction, and improving the management efficiency of the entire life cycle of logistics warehousing. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] 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 or the description of the prior art 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.

[0037] Figure 1 A flowchart of the steps of a logistics warehousing full life cycle management method based on RFID radio frequency identification according to the present invention;

[0038] Figure 2 This is a characteristic relationship flow chart of a logistics warehousing full life cycle management method based on RFID radio frequency identification of the present invention. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a RFID radio frequency identification logistics warehousing full life cycle management method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0041] The following is a detailed description of a specific scheme of a logistics warehousing full life cycle management method based on RFID radio frequency identification provided by the present invention in conjunction with the accompanying drawings.

[0042] See also Figure 1 , which shows a flowchart of a method for managing the entire life cycle of logistics warehousing using RFID radio frequency identification provided by an embodiment of the present invention, the method comprising the following steps:

[0043] Step S001: Collect the inventory data of each type of goods at each recording time.

[0044] It should be noted that when predicting the inventory of various types of items in a logistics warehousing system, existing methods usually use the UKF (Unscented Kalman Filter) unscented Kalman filter algorithm to predict the inventory of each type of item; however, due to the different external environmental noise conditions of items in different locations in the storage environment, the traditional UKF unscented Kalman filter algorithm has a large number of redundant sigma point vectors, which reduces the accuracy of data prediction and thus reduces the management efficiency of the entire life cycle of logistics warehousing.

[0045] Specifically, first, inventory data needs to be collected. The specific process is: install an RFID radio frequency tag on each item in the warehouse, install an RFID scanning device on each shelf, use the RFID scanning device to scan the quantity of goods in the warehouse once every 5 minutes as a recording moment, and transmit the scanned inventory data to the goods database of the central computer. A number of goods types T1 is preset. This embodiment is described by taking T1=5 as an example. This embodiment is not specifically limited, and T1 can be determined according to the specific implementation situation; obtain the inventory data of T1 goods types in the goods database for the past week.

[0046] So far, the inventory data of all types of goods at each recording time are obtained through the above method.

[0047] Step S002: According to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, the selection degree of each sigma point vector is obtained; according to the selection degree of each sigma point vector, all sigma point vectors are divided to obtain a number of warehouse sigma point vectors at each recording moment.

[0048] It should be noted that the UKF unscented Kalman filter algorithm is a nonlinear filtering algorithm that approximates the propagation and measurement update process of nonlinear functions by selecting a set of sigma point vectors to be selected; and the sigma point vector is formed based on the state mean around the current moment, representing the estimated possible range of state changes under the current state. By propagating nonlinear functions and mapping observation models on these sigma point vectors, the propagation results and predicted values ​​of nonlinear functions under the current state estimation can be obtained, thereby realizing the prediction and update of the corresponding inventory data of logistics warehousing, and then realizing the data management of the entire life cycle of logistics warehousing.

[0049] It should be further explained that, in the actual warehouse environment, the RFID scanning devices in different warehouse locations will be affected by noise interference from different environments, and the degree of noise interference in different environments will change to varying degrees with the number of goods in the warehouse, the movement of personnel, and electronic equipment, resulting in a large number of sigma point vectors obtained by the traditional UKF unscented Kalman filter algorithm. There are many sigma point vectors that cannot accurately represent the state of the corresponding recording time. For the same type of goods, although the corresponding inventory data collected for the type of goods is greatly affected by external noise, the corresponding actual inventory will change continuously over time, and there is a certain degree of data trend change that conforms to the actual situation as a whole; therefore, in order to improve the efficiency of logistics warehousing full life cycle management, this embodiment analyzes the change trend of inventory data of each type of goods in the historical time corresponding to different recording moments, and the correlation between different sigma point vectors and corresponding state vectors, and selects several storage sigma point vectors at each recording moment for subsequent data management.

[0050] Furthermore, taking any one recording moment as an example, each recording moment before the recording moment is recorded as a historical comparison recording moment; taking any one historical comparison recording moment and any one type of goods as an example, the absolute value of the difference in inventory data of the type of goods between the recording moment and the historical comparison recording moment is recorded as the inventory change value between the recording moment and the historical comparison recording moment; the inventory change values ​​between the recording moment and all historical comparison recording moments are linearly normalized, and each normalized inventory change value is recorded as an inventory change factor; the maximum value of the inventory change factors between the recording moment and all historical comparison recording moments is recorded as the historical inventory limit change factor of the type of goods at the recording moment.

[0051] Furthermore, according to the overall stability of the inventory data of the commodity type at all historical comparison recording moments before the recording moment, the historical inventory stability of the commodity type at the recording moment is obtained. As an example, the historical inventory stability of the commodity type at the recording moment can be calculated by the following formula:

[0052]

[0053] In the formula, α represents the historical inventory stability of the commodity type at the time of recording; N represents the number of all historical comparison recording times at the time of recording; b n Indicates the inventory data of the goods type at the nth historical comparison record time; It represents the mean of the inventory data of the commodity type at all historical comparison recording moments; σ represents the standard deviation of the inventory data of the commodity type at all historical comparison recording moments. The greater the historical inventory stability of the commodity type at the recording moment, the more stable the inventory quantity change of the commodity type in the historical time before the recording moment.

[0054] Further, the product of the historical inventory limit change factor and the historical inventory stability is recorded as the type data trend of the commodity type at the recording time; the type data trend of all commodity types at the recording time is obtained. A state vector is constructed according to the inventory data of each commodity type at the recording time, and the state vector is recorded as the moment commodity state vector of the recording time; a state vector is constructed according to the type data trend of all commodity types at the recording time, and the state vector is recorded as the moment commodity trend vector of the recording time; the vector after the moment commodity trend vector is transposed is recorded as the moment commodity comparison vector; the state transfer matrix of the recording time is obtained according to the moment commodity state vector of the recording time; the vector obtained by multiplying the moment commodity comparison vector and the state transfer matrix is ​​recorded as the commodity comparison vector of the recording time. The moment commodity state vector contains the inventory data of multiple commodity types, and the state vector is a column vector; in addition, the process of constructing the state vector according to the data and the process of obtaining the state transfer matrix according to the state vector are both well-known contents of the UKF (Unscented Kalman Filter) unscented Kalman filter algorithm, which will not be repeated in this embodiment.

[0055] Furthermore, the inventory data of all types of goods from the first recording moment to the recording moment, as well as the state vector of the goods at the recording moment, are input into the UKF unscented Kalman filter algorithm to obtain several sigma point vectors at the recording moment; taking any sigma point vector as an example, the cosine similarity between the goods comparison vector and the sigma point vector is recorded as the goods inventory similarity between the goods comparison vector and the sigma point vector. The sigma point vector is a column vector; in addition, the process of obtaining the sigma point vector according to the data and the state vector is a well-known content of the UKF (Unscented Kalman Filter) unscented Kalman filter algorithm, which will not be repeated in this embodiment.

[0056] Furthermore, the last historical comparison recording time before the recording time is recorded as the historical nearest comparison time, and the selection degree of the sigma point vector is obtained according to the change difference of the inventory data of all goods types between the recording time and the historical nearest comparison time, and the similarity of the goods inventory between the goods comparison vector and the sigma point vector. As an example, the selection degree of the sigma point vector can be calculated by the following formula:

[0057]

[0058] In the formula, γ represents the selection degree of the sigma point vector; τ represents the similarity of the goods inventory between the goods comparison vector and the sigma point vector; M represents the number of all goods types; b1 mb2 represents the inventory data of the mth commodity type at the time of recording; m represents the inventory data of the mth commodity type at the historical nearest neighbor comparison time; α m Indicates the historical inventory stability of the mth commodity type at the time of recording; || indicates taking the absolute value. The greater the selection degree of the sigma point vector, the more fully the data change characteristics of the sigma point vector when the inventory data of each commodity type changes abnormally, reflecting that the sigma point vector should be selected.

[0059] Furthermore, the sequence of all sigma point vectors arranged in descending order according to the selection degree is recorded as the sigma sequence to be selected. Taking any two adjacent sigma point vectors in the sigma sequence as an example, the absolute value of the difference in the selection degree between the two sigma point vectors is recorded as the division degree of the two sigma point vectors; among the division degrees of all any two adjacent sigma point vectors in the sigma sequence to be selected, the two sigma point vectors corresponding to the smallest division degree are recorded as the sigma point vectors to be divided; among the two sigma point vectors to be divided, the second sigma point vector to be divided is recorded as the termination sigma point vector; each sigma point vector between the first sigma point vector and the termination sigma point vector in the sigma sequence to be selected is recorded as the storage sigma point vector at the recording moment. A number of storage sigma point vectors at all recording moments are obtained.

[0060] At this point, several storage sigma point vectors of all recording moments are obtained through the above method.

[0061] Step S003: Perform data prediction management based on several storage sigma point vectors at each recording moment.

[0062] Specifically, the storage sigma point vector at each recording moment is used as the sigma point vector actually used at each recording moment. According to the inventory data of each type of goods at each recording moment and the sigma point vector actually used at each recording moment, the prediction value of each type of goods at each recording moment is obtained through the UKF unscented Kalman filter algorithm. The prediction values ​​of all types of goods at each recording moment are input into the AnalyticDB database of the central computer to complete the full life cycle management of logistics warehousing. The AnalyticDB database is a real-time analysis database. In addition, the process of obtaining the prediction value based on the data and the sigma point vector is a well-known content of the UKF unscented Kalman filter algorithm, which will not be repeated in this embodiment.

[0063] At this point, this embodiment is complete. Figure 2, which shows a characteristic relationship flow chart of a logistics warehousing full life cycle management method based on RFID radio frequency identification.

[0064] 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 principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A logistics warehousing full life cycle management method based on RFID radio frequency identification, characterized in that: The method comprises the following steps: Collect inventory data for each type of goods at each recording moment; According to the inventory data of each type of goods at each recording moment, several sigma point vectors at each recording moment are obtained through the UKF unscented Kalman filter algorithm; any recording moment is recorded as the target recording moment, and each recording moment before the target recording moment is recorded as the historical comparison recording moment; according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, the selection degree of each sigma point vector is obtained; according to the selection degree of each sigma point vector, all sigma point vectors are divided to obtain several storage sigma point vectors at each recording moment; The selection degree of each sigma point vector is obtained according to the overall trend stability of the inventory data of each type of goods at all historical comparison recording moments, and the correlation between each sigma point vector at the target recording moment and the inventory data at different historical recording moments, including the specific method as follows: For any historical comparison recording time and any commodity type, according to the difference in inventory data between the target recording time and the historical comparison recording time, the historical inventory limit change factor of the commodity type at the target recording time is obtained; According to the overall change trend of the inventory data of the goods type at all historical comparison recording moments before the target recording moment, the historical inventory stability of the goods type at the target recording moment is obtained; According to the historical inventory limit change factor of the goods type at the target recording time and the correlation of the historical inventory stability, a vector combination is constructed to obtain the goods comparison vector at the target recording time; For any sigma point vector at the target recording time, the similarity between the cargo comparison vector and the sigma point vector at the target recording time is obtained; The last historical comparison record time before the target record time is recorded as the historical nearest comparison time. The selection degree of the sigma point vector is obtained according to the change difference of the inventory data of all goods types between the target record time and the historical nearest comparison time, and the similarity of the goods inventory between the goods comparison vector and the sigma point vector. The specific method of obtaining the historical inventory stability of the goods type at the target recording time according to the overall change trend of the inventory data of all historical comparison recording times before the target recording time includes: ； In the formula, Indicates the historical inventory stability of the goods type at the target recording time; The number of all historical comparison record moments representing the target record moment; Indicates the type of goods in Inventory data at the time of historical comparison and recording; It represents the average inventory data of the commodity type at all historical comparison record times; It represents the standard deviation of the inventory data of the commodity type at all historical comparison record moments; The method of obtaining the selection degree of the sigma point vector according to the change difference of the inventory data of all goods types between the target recording time and the historical neighbor comparison time, and the goods inventory similarity between the goods comparison vector and the sigma point vector, includes the following specific methods: ； In the formula, Indicates the selection degree of the sigma point vector; Represents the similarity of goods inventory between the goods comparison vector and the sigma point vector; Indicates the quantity of all goods types; Indicates Inventory data of each type of goods at the target recording time; Indicates Inventory data of each type of goods at the historical nearest comparison time; Indicates The historical inventory stability of each product category at the target recording time; Indicates taking the absolute value; Data prediction management is performed based on several stored sigma point vectors at each recording moment.

2. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 1, it is characterized by: The method of obtaining the historical inventory limit change factor of the commodity type at the target recording time according to the difference in inventory data between the target recording time and the historical comparison recording time includes the following specific methods: The absolute value of the difference between the inventory data of the target recording time and the historical comparison recording time of the commodity type is recorded as the inventory change value between the target recording time and the historical comparison recording time; Linearly normalize the inventory change values ​​at the target recording time and all historical comparison recording times, and record each normalized inventory change value as the inventory change factor; The maximum value of the inventory change factors between the target recording time and all historical comparison recording times is recorded as the historical inventory limit change factor of the commodity type at the target recording time.

3. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 1, it is characterized in that: The method of constructing a vector combination based on the historical inventory limit change factor of the goods type at the target recording time and the correlation between the historical inventory stability and the target recording time to obtain the goods comparison vector includes the following specific methods: According to the historical inventory limit change factor and historical inventory stability of the goods type at the target recording time, the moment goods trend vector at the target recording time is obtained; the vector after the moment goods trend vector is transposed is recorded as the moment goods comparison vector; According to the inventory data of each type of goods at the target recording time, the state vector is constructed through the UKF unscented Kalman filter algorithm, and the state vector is recorded as the moment-by-moment goods state vector of the target recording time; according to the moment-by-moment goods state vector of the target recording time, the state transfer matrix of the target recording time is obtained through the UKF unscented Kalman filter algorithm; the vector obtained by multiplying the moment-by-moment goods comparison vector and the state transfer matrix is ​​recorded as the goods comparison vector of the target recording time.

4. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 3, it is characterized by: The specific method of obtaining the cargo trend vector at the target recording time according to the historical inventory limit change factor and the historical inventory stability of the cargo type at the target recording time is as follows: The product of the historical inventory limit change factor and the historical inventory stability is recorded as the type data trend of the goods type at the target recording time; the type data trends of all goods types at the target recording time are obtained; according to the type data trends of all goods types at the target recording time, a state vector is constructed through the UKF unscented Kalman filter algorithm, and the state vector is recorded as the momentary goods trend vector at the target recording time.

5. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 1, it is characterized by: The method of obtaining the similarity between the cargo comparison vector and the sigma point vector according to the similarity between the cargo comparison vector and the sigma point vector at the target recording time includes: The cosine similarity between the cargo comparison vector and the sigma point vector is recorded as the cargo inventory similarity between the cargo comparison vector and the sigma point vector.

6. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 1, it is characterized by: The method of dividing all sigma point vectors according to the selection degree of each sigma point vector to obtain a number of storage sigma point vectors at each recording moment includes: For any recording moment, the sequence of all sigma point vectors at the recording moment arranged in descending order according to the selection degree is recorded as the sigma sequence to be selected. For any two adjacent sigma point vectors in the sigma sequence, the absolute value of the difference in the selection degree between the two sigma point vectors is recorded as the division degree of the two sigma point vectors; among the division degrees of all any two adjacent sigma point vectors in the sigma sequence to be selected, the two sigma point vectors corresponding to the smallest division degree are recorded as the sigma point vectors to be divided; among the two sigma point vectors to be divided, the second sigma point vector to be divided is recorded as the termination sigma point vector; each sigma point vector between the first sigma point vector and the termination sigma point vector in the sigma sequence to be selected is recorded as the storage sigma point vector at the recording moment.

7. According to the RFID radio frequency identification logistics warehousing full life cycle management method of claim 1, it is characterized by: The data prediction management is performed according to a number of storage sigma point vectors at each recording moment, including the specific method as follows: The storage sigma point vector at each recording moment is used as the sigma point vector actually used at each recording moment. According to the inventory data of each type of goods at each recording moment and the sigma point vector actually used at each recording moment, the predicted value of each type of goods at each recording moment is obtained through the UKF unscented Kalman filter algorithm. The predicted values ​​of all types of goods at each recording moment are input into the AnalyticDB database of the central computer to complete the full life cycle management of logistics warehousing.

Images