Intelligent shelf management system and application method

Abstract

This invention relates to an intelligent shelf management system and its application method, comprising a single-layer shelf, at least three high-precision weight sensors, and a main control computer. The system achieves precise management through a product loading and registration process and a shelf status monitoring process. During loading and registration, the product barcode is scanned, and the main control computer guides the placement location. After the product is placed, the system records the resulting increase in overall shelf weight and the change in center of gravity, establishing an archive containing product identity, location, weight, and center of gravity shift data. During status monitoring, by continuously monitoring the total weight and center of gravity of the shelf, when an abnormal weight reduction is detected, the system intelligently matches the recorded weight reduction value with the center of gravity change data in the archive to accurately identify the removed product. This invention achieves real-time, accurate single-item-level identification and management of high-value products, with advantages such as low cost, accurate identification, anti-theft, and good inventory management.

Description

Technical Field

[0001] This invention relates to the field of smart supermarket technology, and more specifically, to a smart shelf management system and its application method. Background Technology

[0002] In retail settings such as large supermarkets and chain convenience stores, high-value goods such as tobacco and alcohol, high-end skincare products, and imported snacks are important sources of profit, but they are also the hardest hit areas for merchandise loss (mainly theft); traditional management methods for such goods have significant drawbacks:

[0003] First, it relies on manual supervision; arranging dedicated personnel to be on duty or frequent patrols is costly in terms of manpower, and there are blind spots and fatigue and negligence, so the anti-theft effect is limited.

[0004] Secondly, existing electronic anti-theft measures are ineffective; the hard tags of commonly used electronic article protection (EAS) systems are easily removed or blocked, while soft tags may affect the product packaging, and each tag needs to be demagnetized at checkout, which affects the customer experience; ordinary video surveillance systems are mainly used for post-event traceability, have weak real-time early warning capabilities, and are difficult to bind to specific products.

[0005] Furthermore, inventory management is inefficient; currently, the common practice is to use periodic inventory counts or back-calculation based on point-of-sale (POS) data, which makes it impossible to know the real-time and accurate quantity of goods on the shelves. This leads to two problems: first, untimely replenishment may result in missed sales opportunities; second, it is impossible to detect abnormal inventory reductions (such as theft) in real time, resulting in property losses.

[0006] In recent years, some smart shelf solutions have emerged, such as gravity-sensing shelves with pressure sensors installed under each shelf location. However, such solutions are usually expensive, and a single pressure sensor has the drawback of "knowing the weight but not the item," requiring the use of cameras for monitoring. Moreover, when multiple items of similar weight are placed on the shelf (such as cigarettes and liquor of different brands but with similar weights), the system cannot accurately determine which item was taken based solely on weight changes, significantly reducing its anti-theft and accurate inventory management functions. Furthermore, when multiple customers select items simultaneously, the system struggles to differentiate between them, easily leading to misjudgments.

[0007] Therefore, for the refined management of high-value goods in smart supermarkets, there is an urgent need for an intelligent management solution that can identify the picking and putting of specific items in real time at low cost and with high accuracy, and can effectively cope with complex shopping scenarios. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide an intelligent shelf management system and an application method of the intelligent shelf management system, in view of the above-mentioned defects of the prior art.

[0009] The technical solution adopted by this invention to solve its technical problem is:

[0010] An intelligent shelf management system is constructed, which includes a single-layer shelf for displaying goods. At least three high-precision weight sensors are installed on the bottom load-bearing structure of the shelf. A main control computer is connected to these sensors and is responsible for collecting and processing sensor data. A set of management software runs in the main control computer, which mainly executes two core processes: a goods loading and registration process and a shelf status monitoring process.

[0011] The steps of the product registration process are as follows: Staff use a barcode scanner to scan the barcode of the product to be shelved to obtain its unique identification code; the main control computer assigns a designated placement position (represented by "row number, column number") to the product based on a pre-stored digital map of the shelf, and sends this position information to the staff's handheld device as a guide; after the staff places the product in the designated position according to the guide, the main control computer immediately obtains data from the sensors to calculate how much weight the shelf has gained overall, and how far and in which direction the shelf's center of gravity has shifted; subsequently, the system bundles the product's identification code, placement position, increased weight, and resulting center of gravity shift data together to form a complete digital file, which is then stored in a dedicated database; this process establishes a unique "physical characteristic ID card" for each product.

[0012] The shelf status monitoring process involves the following steps: The main control computer continuously reads data from the sensors, monitoring the total weight and center of gravity of the shelf in real time. Once the system detects that the total weight has decreased by more than a preset threshold in a very short period of time (e.g., a decrease of more than 50 grams, excluding human interference), it immediately determines that an item has been taken away and records the weight reduction and the change in the center of gravity at that moment. Then, the system uses these two key data points as "clues" to perform a comprehensive comparison and intelligent search in the previously established product file database. The goal is to find a file record whose recorded weight and theoretical center of gravity change highly match the currently monitored "clues." Once found, the system outputs the product identification code corresponding to that file, thereby accurately identifying which product was taken away.

[0013] This invention provides a product loading and registration method for the above-mentioned system; the method includes the following sequential steps: scanning the product barcode to obtain the code; the system allocates vacant storage locations (row and column coordinates) according to the electronic map and issues them to the staff; the staff places the products according to the map; the system detects and records the weight increment and center of gravity offset vector after placement; finally, the product code, location, weight, and offset vector are bound together to create a new digital file for storage.

[0014] This invention provides a product retrieval sensing method applied to the above-mentioned system; the method includes the following ordered steps: continuously and frequently monitoring the total weight and center of gravity of the shelf; triggering an event when a sudden decrease in weight is detected, recording the weight loss value and the center of gravity change vector; immediately searching for matching product records in a digital archive; and finally outputting the identification code of the successfully matched product.

[0015] Preferably, in the product loading and registration process, the system automatically allocates empty storage locations by querying the electronic shelf layout map and sends the specific row and column numbers to the staff's handheld devices to achieve guided loading and ensure the absolute accuracy of product placement, which is the basis for subsequent accurate identification.

[0016] Preferably, in the shelf status monitoring process, the system calculates the actual direction and distance of movement based on the monitored changes in the center of gravity. Then, it iterates through each product file in the database and, based on the product position and weight recorded in the file, reverse-engineers "theoretically, if this product were taken away, the direction and distance of the center of gravity movement would be affected." Next, the system filters out files that simultaneously meet two stringent conditions: first, the weight recorded in the file is almost equal to the weight reduction value monitored this time (e.g., the error is within a few grams); second, the calculated theoretical direction of center of gravity movement is basically consistent with the actual direction (e.g., the angle is within a very small degree), and the proportion of the movement distance is within a reasonable range. Through this double verification, the accuracy of matching is greatly improved.

[0017] Preferably, when the above matching conditions filter out more than one candidate product (which may happen in a very few cases where the weight and location are very similar), the system will initiate a secondary judgment logic to select the product whose registered weight is closest to the reduced weight detected this time as the final result, thereby solving the fuzzy matching problem.

[0018] Preferably, there are four weight sensors, which are fixedly installed at the four corners of the bottom of the rectangular shelf. This four-point support layout can most stably and reliably calculate the two-dimensional plane center of gravity coordinates of the entire shelf, which is the physical basis for the operation of the whole system.

[0019] Preferably, the main control computer has the intelligent capability to handle multiple customers picking up goods simultaneously (concurrent events). The logic is as follows: when the system detects multiple consecutive weight decreases within a very short time window (e.g., 5 seconds), it merges these into a "concurrent picking up event package." The system calculates the total weight reduction of this event package and analyzes the continuous trajectory of the center of gravity change during this period. Then, the system searches the product database for a "combination package" consisting of multiple items. The total weight of this "combination package" matches the detected total weight reduction, and when the items in this "combination package" are simulated to be picked up in a certain order, the calculated theoretical trajectory of the center of gravity change is most similar to the actual monitored trajectory. Finally, the system identifies all the items in this "combination package" as items picked up simultaneously. This mechanism effectively solves the common problem of concurrent operation identification in the field of intelligent shelves.

[0020] Preferably, when allocating product placement locations, the system adopts an intelligent recommendation strategy: if the currently stocked product is already displayed on the shelf, the system prioritizes allocating an empty space adjacent to the existing product for centralized display and management; if there is no adjacent empty space, it allocates a relatively nearest empty space; this not only conforms to retail display principles, but also makes the center of gravity distribution of similar products relatively concentrated, which is beneficial for subsequent analysis.

[0021] The beneficial effects of this invention are as follows:

[0022] 1. Precise and effective anti-theft: It can identify the identity of each high-value item that is taken in real time and accurately; once an item is taken and the settlement is not completed within the specified time, the system can immediately alarm and clearly indicate what item is missing, which greatly improves the targeting and timeliness of anti-theft.

[0023] 2. Real-time transparent inventory: The digital archive established by the system is a real-time and accurate inventory list for the shelves; without the need for manual inventory checks, managers can keep track of the status of every valuable item on the shelves at any time, achieving precise inventory management at the "single item level" and completely eliminating replenishment delays or backlogs caused by inaccurate data.

[0024] 3. High recognition accuracy and strong anti-interference: The system adopts a dual-factor matching mode of "weight + center of gravity shift", which has an extremely low false judgment rate compared with single gravity sensing. Even for different products with very similar weights, the system can still effectively distinguish them because their positions on the shelf result in different directions of center of gravity shift.

[0025] 4. Significant cost advantages: Only a few sensors need to be installed at the bottom of the entire shelf. There is no need to attach RFID tags to each product or modify each shelf location. The overall hardware cost and deployment difficulty are far lower than other existing smart shelf solutions, making it particularly suitable for intelligent upgrades and transformations based on existing supermarket shelves.

[0026] 5. High level of intelligence: Guided stocking reduces human error; automatic out-of-stock identification can directly link with the replenishment system; powerful concurrent event handling capabilities adapt to the real supermarket shopping environment and are highly practical. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort:

[0028] Figure 1 This is a schematic diagram of the intelligent shelf management system according to a preferred embodiment of the present invention;

[0029] Figure 2 This is a flowchart of the loading and registration process according to a preferred embodiment of the present invention;

[0030] Figure 3 This is a flowchart of the pickup monitoring and identification process according to a preferred embodiment of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, a clear and complete description will be provided below in conjunction with the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0032] like Figure 1 As shown, see also Figure 2 and Figure 3 Cigarettes, alcohol, and cosmetics are used as examples to illustrate this:

[0033] Example 1: System Application and Deployment in High-End Cigarette Counters

[0034] A large supermarket deployed the system of this invention in its closed-loop cigarette counter, which is an independent single-layer glass shelf 1.

[0035] Hardware deployment: Underneath the four support columns 10 at the bottom of the shelf, a high-precision weighing sensor 2 with a range of 50 kg and an accuracy of ±2 g is installed in each column. The four sensors are connected to an industrial control box 3 (main control computer) hidden inside the cabinet. A tablet computer 4 with barcode scanning function is provided outside the cabinet for store staff to use.

[0036] Inventory registration process: The clerk retrieves a carton of "Chunghwa (Hard) Cigarettes" (standard weight approximately 280 grams) from the warehouse; he uses a tablet to scan the barcode on the outer box of the cigarettes, and the system enters the product code "ZJ_ZH_YI";

[0037] The main control computer queries the shelf layout and finds that the same brand of cigarettes is already displayed in the 3rd and 4th columns of the 1st row. It then automatically allocates the adjacent empty space in the 5th column of the 1st row and displays "Location: 1st row, 5th column" on the tablet.

[0038] The clerk placed the cigarettes in the designated location; 2 seconds later, the system detected that the total weight of the shelf had increased by 281 grams, and the center of gravity had shifted a certain distance to the left rear of the shelf.

[0039] The system then creates a file: [Code: ZJ_ZH_YI, Position: 1-5, Actual Weight: 281 grams, Center of Gravity Offset Data: (X1, Y1)];

[0040] In this way, each pack of cigarettes has a unique physical identification file.

[0041] Pickup identification and anti-theft: A customer, under the supervision of a store clerk, indicated that he / she wanted to buy a pack of "Zhonghua (Hard)" cigarettes;

[0042] He took the cigarette from the 1st row, 5th column; the system instantly detected a weight reduction of 281 grams, indicating a reverse shift in the center of gravity;

[0043] The system immediately matched the data in the database, instantly identified the retrieved item as "ZJ_ZH_YI", and displayed a message on the screen to the clerk: "Item ZJ_ZH_YI has been retrieved."

[0044] The clerk then scanned the pack of cigarettes on the POS machine to complete the transaction; upon receiving the transaction information, the system automatically updated the status of the corresponding product in the database to "sold," and the process ended.

[0045] If a customer fails to settle the payment within the set time (e.g., 1 minute) after taking the cigarettes, the system will issue an audible and visual alarm and clearly indicate which location and code of the cigarettes is abnormal, allowing security personnel to intervene immediately.

[0046] Example 2: Handling Concurrent Pickup Events

[0047] In the beverage section, a smart shelf displays a variety of imported spirits;

[0048] Customer A takes a bottle of whiskey weighing 825 grams from the left side of the shelf (position 2,1), and almost simultaneously, customer B takes a bottle of red wine weighing 750 grams from the right side of the shelf (position 2,8).

[0049] System processing: The main control computer detected two weight drops within 3 seconds: the first was about 825 grams, with the center of gravity shifting to the left; then the second was about 750 grams, with the center of gravity shifting significantly to the right.

[0050] The system determined this to be a concurrent event; it calculated that the total weight reduction was approximately 1575 grams and analyzed that the center of gravity completed a "left-then-right" movement trajectory in a short period of time.

[0051] The system initiated a concurrent analysis algorithm to search the database for product combinations with a total weight close to 1575 grams. It tried various combinations and ultimately found that the combination "code W1 (whiskey, 825 grams, position 2,1) + code R1 (red wine, 750 grams, position 2,8)" simulated a center of gravity movement trajectory in the order of "taking W1 first, then R1," which highly matched the detected "left then right" trajectory. Other combinations with a total weight of 1575 grams could not be matched in their simulated trajectories.

[0052] Therefore, the system accurately outputs the identification result: Items W1 and R1 have been taken. The inventory has been updated correctly.

[0053] Example 3: Intelligent Replenishment Trigger

[0054] In the cosmetics section, there are 5 bottles of a certain serum (code JH01, registered weight 152g / bottle) on the shelf, located in different positions. After the system identifies that the product has been taken and paid for three consecutive times through the above process, the number of bottles of the product in the database that are "on the shelf" changes to 2.

[0055] The system's built-in replenishment rules automatically generate a replenishment reminder when the quantity of a product on the shelf falls below 3 bottles. The main control computer then immediately generates a replenishment task in the background: "Product JH01, requires replenishment of 3 bottles, recommended replenishment locations: 3-4, 3-5, 3-6," and sends it to the warehouse manager's mobile terminal. Warehouse staff can then use this information to precisely prepare and stock inventory, achieving closed-loop automation of inventory management.

[0056] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. An intelligent shelf management system, characterized in that, include: A single-layer shelf with at least three sensors installed at the bottom for sensing weight; The main control computer is electrically connected to the sensor and is used to receive data from the sensor and run management programs. The management procedures include: a merchandise stocking registration process and a shelf status monitoring process; The product placement and registration process includes the following steps: scanning the barcode of the product to be placed on the shelf with a barcode scanner to obtain the product's unique number; the main control computer, based on a stored shelf layout diagram, sends a target shelf location information containing the specific number of layers, rows, and columns to the handheld device carried by the stock clerk; after the stock clerk places the product in the target shelf location, the main control computer obtains the overall shelf weight increase and overall shelf center of gravity change values ​​reported by the sensors; the main control computer associates the product's unique number, the target shelf location information, the weight increase value, and the center of gravity change value, and saves them as a product file record in its local archive database. The shelf status monitoring process includes the following steps: the main control computer continuously reads the current total weight of the shelf and the current center of gravity position of the shelf reported by the sensors; when it is found that the current total weight of the shelf decreases by more than a preset alarm weight value within a set time, the current weight decrease value and the current change value of the shelf center of gravity position are recorded; the main control computer compares and searches the file database based on the weight decrease value and the change value of the shelf center of gravity position, finds the matching product file record, and outputs the unique product number recorded in the record.

2. The intelligent shelf management system according to claim 1, characterized in that, In the aforementioned goods loading and registration process, the specific steps for sending target storage location information to the handheld device are as follows: After scanning the product barcode, the main control computer finds an empty shelf location that has not yet been stocked from an electronic layout diagram that corresponds to the actual shelf structure, and sends the row and column number information of this empty shelf location to the handheld device.

3. The intelligent shelf management system according to claim 1, characterized in that, The specific steps for comparing and searching the archive database in the shelf status monitoring process include: Calculate the direction and distance of the center of gravity movement represented by the change in the center of gravity position of the shelf; Examine each product record in the database one by one; For each product record, based on the actual placement position corresponding to the number of rows and columns recorded, as well as the recorded weight increase, the theoretical direction and distance of the center of gravity shift that would result from removing the product from the shelf are estimated. Select product profile records that meet both of the following criteria: The first standard states that the absolute value of the difference between the recorded increase in weight and the recorded decrease in weight is less than 5 grams. The second criterion is that the angle difference between the estimated theoretical center of gravity movement direction and the calculated center of gravity movement direction is less than 10 degrees, and the ratio of the estimated theoretical center of gravity movement distance to the calculated movement distance is between 0.8 and 1.

2.

4. The intelligent shelf management system according to claim 3, characterized in that, When more than one product file record that simultaneously meets the two criteria is selected, the main control computer is further configured to: select the record from these records whose absolute value of the difference between the recorded weight increase value and the current monitored weight decrease value is the smallest, and output its unique product number.

5. The intelligent shelf management system according to claim 1, characterized in that, There are four sensors, which are fixedly installed at the four corners of the bottom of the shelf.

6. The intelligent shelf management system according to claim 1, characterized in that, The main control computer is configured to handle situations where multiple customers pick up their goods simultaneously, specifically including: If the total weight of the shelf is continuously monitored to decrease multiple times within a preset time window, it is determined to be a concurrent picking event; The total weight reduction value of the multiple descent events is calculated cumulatively, and a continuous data sequence of changes in the center of gravity position is obtained; Find a product combination from the archive database such that the difference between the sum of the registered weights of all products in the combination and the total weight reduction is within an allowable range, and the theoretical center of gravity change sequence generated when the product combination is removed in sequence matches the obtained center of gravity position change data sequence to the highest degree. Output the unique IDs of all items in this product combination.

7. A method for applying an intelligent shelf management system, applicable to the intelligent shelf management system as described in any one of claims 1-6, characterized in that, The method includes the following steps: Step 1: The stock clerk uses a barcode scanner to scan the barcode on the outer packaging of the first item to be put on the shelf, and the main control computer obtains the item code of this item; Step 2: The main control computer assigns an empty physical placement position to this product based on the pre-set electronic shelf layout diagram. This placement position is uniquely determined by the number of rows and columns. Step 3: The main control computer sends the specific location information, including the number of rows and columns, to the smart terminal held by the stock clerk. Step 4: The stock clerk checks the prompts on the smart terminal and places the first item in the corresponding row and column location on the shelf; Step 5: After the goods are placed securely, the main control computer receives data from the sensors, calculates the first increase in weight of the entire shelf, and the first offset vector of the shelf's center of gravity moving from its original position to its new position. Step 6: The master computer adds a new record to the local archive. This record must contain at least the product code, number of rows, number of columns, first weight value, and first offset vector.

8. The application method of the intelligent shelf management system according to claim 7, characterized in that, The specific rules for allocating vacant storage spaces for goods are as follows: when some goods with the same product code are already placed on the shelf, priority is given to allocating vacant storage spaces that are adjacent to the existing goods; if there are no adjacent vacant storage spaces, the nearest vacant storage space is selected from the electronic layout diagram for allocation.

9. A method for applying an intelligent shelf management system, used in any one of claims 1-6, characterized in that, The method includes the following steps: Monitoring steps: The main control computer reads the data from the sensors continuously at a high frequency, and calculates and updates the total weight and center of gravity coordinates of the shelf at this moment in real time; Triggering steps: When it is found that the total weight at this moment has decreased by more than a set threshold weight in a single sampling instant, a picking event is triggered, and the weight reduction value and the change vector of the center of gravity position coordinates caused by this event are immediately recorded. Matching step: The main control computer immediately searches the product file database for the product file record that best matches the weight reduction value and the center of gravity position coordinate change vector that were just recorded; Output steps: Extract the product code from the matched product file record as the identification conclusion of the product taken from the shelf this time.

10. The application method of the intelligent shelf management system according to claim 9, characterized in that, The specific operation process of the matching step includes: Based on the recorded change vector of the center of gravity position coordinates, calculate the actual offset direction and the actual offset distance; Traverse the product records in the archive database, and for each record, calculate the theoretical center of gravity offset vector and its corresponding theoretical direction and theoretical distance based on its position and weight; Filter out product records that simultaneously meet the following conditions: the difference between its registered weight and the event weight reduction value is within the first tolerance range, the deviation between its theoretical direction and actual direction is within the second tolerance range, and the ratio of its theoretical distance to actual distance is within the third tolerance range. The final matching product records are determined based on the filtering results.

Images