Perception card matching method for warehouse supervision

Abstract

The application discloses a kind of perception card matching method for warehouse supervision, comprising the following steps: S1, supervisory equipment lifting process;S2, motion matching decision;S3, position, weight matching decision;S4, the calculation of decision fusion strategy probability;S5, filtering algorithm;S6, dynamic and static detection algorithm;The present application uses the perception card system based on IMU to replace the traditional way, using the IMU in the perception card and other sensors, which can effectively detect the steel coil lifting process.And can and supervisory equipment, i.e. pre-installed with related equipment operating machine, including travelling crane, forklift, etc., automatically establish matching relationship.Thus realize the automatic, effective supervision of steel coil and other movable property, provide technical possibility for movable collateral, movable securitization.

Description

Technical Field

[0001] This invention belongs to the field of warehouse supervision technology, specifically relating to a sensor card matching method for warehouse supervision. Background Technology

[0002] In the bulk steel coil logistics industry, overhead cranes and other machinery are frequently used for lifting, loading, unloading, and transferring steel coils. Traditionally, this involves supervisors manually recording data for management and oversight, which suffers from low automation, susceptibility to errors, limited information, and inherent supervisory risks.

[0003] To address this issue, we propose an IMU-based sensing card system to replace traditional methods. Utilizing sensors such as the IMU within the sensing card, the steel coil hoisting process can be effectively detected. Furthermore, it can automatically establish a matching relationship with monitoring equipment—operating machinery pre-installed with relevant equipment, including overhead cranes and forklifts. This sensing card matching method for warehouse supervision, enabling automatic and effective monitoring of movable assets such as steel coils, solves the aforementioned problems. Summary of the Invention

[0004] The purpose of this invention is to provide a sensor card matching method for warehouse supervision, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a sensor card matching method for warehouse supervision, comprising the following steps: S1. Lifting Process of Monitoring Equipment: The complete lifting process of monitoring equipment includes the following steps: moving the mobile device to the lifting position - colliding with the cargo during equipment connection - starting lifting - lifting completed - moving the mobile device to the lowering position - starting lowering - lowering completed - equipment returns; a matching association is established between the equipment and the sensing card to identify the cargo information lifted by the monitoring equipment. The establishment method is to make motion time decisions, weight decisions, and position decisions based on the data messages uploaded by the sensing card and the weight and position devices, and then perform the fusion of the three decisions to obtain the decision probability of all cargoes. The cargo with the highest probability is selected as the matching cargo, triggering relevant record management operations for that cargo and relevant alarms for unauthorized lifting. S2. Motion Matching Decision: Matching decisions are made using the motion characteristics of the monitoring equipment and the cargo. The motion information of the monitoring equipment is provided by the weight and position device, and the motion information of the cargo is provided by the sensing card. The motion information used for motion matching includes motion direction, motion speed, and motion displacement. The specific implementation mode is to calculate the total movement duration reported by each sensing card during the hoisting period, landing period and cargo-carrying movement period of the supervision equipment. Due to the requirement of the alarm time limit, the movement duration during the hoisting period of the supervision equipment is used as the matching basis. The hoisting period is subject to the reception of the start hoisting frame, traces back the L0 period forward and tracks the Li period backward. Assuming that the start hoisting time is Tr, the total matching period is (Tr-L0, Tr+Li), and it is required that L1<Li<L2. Within this allowable time range, the system automatically performs multiple matching continuously, and the minimum time interval between two adjacent matchings is l, that is, the multiple matching times are Tr+L1, Tr+L1+l, Tr+L1+2l, … and so on in sequence until Tr+L2 is exceeded, wherein L0 is a parameter representing the length of the period traced back for matching, L1 and L2 are the minimum and maximum matchable duration parameters respectively, and l represents the time interval between two consecutive matchings; The matching period duration L involved in movement time matching is defined as (L0+Li), wherein L0 is a parameter representing the length of the period traced back to before the start of hoisting required for matching; Li is the actual time length from the start of hoisting when the matching operation is actually executed, and L1<Li<L2, where L1 and L2 are the minimum and maximum matchable durations respectively; S3. Position and weight matching decision: both position data and weight data are calculated respectively to obtain the confidence degrees of position and weight according to the position or weight matching decision calculation flow chart; S4. Calculation of decision fusion strategy probability: according to the movement matching decision calculation flow in step S2 and the position and weight matching decision calculation flow in step S3, the movement matching decision confidence SI, the weight matching decision confidence WI and the position matching decision confidence XI are calculated respectively, and then the decision probabilities of each cargo are respectively , the cargo with the maximum decision probability is selected as the matched cargo, if , it is considered that the matching is unsuccessful; S5. Filtering algorithm: the filtering part adopts a mean-median filtering combined with single iteration prediction filtering algorithm, and the prediction filtering algorithm is as follows: wherein R is an algorithm parameter. In the mean-median filtering, the number of median values n and the total length of the filtering window N are algorithm parameters; S6. Dynamic and static detection algorithm: a state machine method is adopted for data smoothing and delay detection, and the state machine flow is as follows: State 0: if feature_var>=THR1 Accumulator 1 If counter 1 is greater than or equal to CNT1, then switch the state to 1 and clear counter 1. else Clear counter 1 State 1: if feature_var<=THR2 Accumulator 2 If counter 2 is greater than or equal to CNT2, then switch the state to 0 and clear counter 2. else Clear counter 2 In the angular velocity variance window WIN and the delay detection filtering time DELAY, the algorithm parameters are used. In the state machine method, THR1, THR2, CNT1, and CNT2 are all algorithm parameters.

[0006] The sensing card has a built-in IMU sensor.

[0007] The weight location device refers to the integrated system device that collects weight location information.

[0008] The technical effects and advantages of this invention are as follows: This sensor card matching method for warehouse supervision requires monitoring goods such as steel coils. To identify the goods lifted by the monitoring equipment, a matching association needs to be established between the equipment and the sensor card. The method involves making motion time decisions, weight decisions, and position decisions based on data uploaded by the sensor card and the weight and position device, and then fusing these three decisions to obtain the decision probability for all goods. The goods with the highest probability are selected as the matching goods, triggering relevant alarms for those goods. Motion matching decision refers to using the motion characteristics of the monitoring equipment and the goods to make matching decisions. The motion information of the monitoring equipment is provided by the weight and position device, and the motion information of the goods is provided by the sensor card. This involves further matching using motion duration information or other motion information, such as motion direction, speed, and displacement. Specifically, the method calculates the motion information reported by the sensor card and the motion information of the monitoring equipment provided by the weight and position device during the lifting, lowering, and moving periods of the monitoring equipment, and executes the matching decision process in different time periods.

[0009] Using an IMU-based sensing card system to replace traditional methods, the sensing card utilizes sensors such as the IMU to effectively detect the steel coil hoisting process. It can also automatically establish a matching relationship with monitoring equipment—operating machinery pre-installed with relevant equipment, including overhead cranes and forklifts. This enables automatic and effective monitoring of movable assets such as steel coils, providing technological possibilities for movable asset pledging and securitization. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the complete process of lifting the monitoring equipment according to the present invention; Figure 2 This is a flowchart illustrating the overall matching process of the present invention; Figure 3 This is a flowchart of the motion matching decision calculation process of the present invention; Figure 4 This is a flowchart of the position or weight matching decision calculation process of the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] This invention provides a sensor card matching method for warehouse supervision, comprising the following steps: S1. Lifting Process of Monitoring Equipment: The complete lifting process of monitoring equipment includes the following steps (e.g., Figure 1 (As shown): Mobile device moves to lifting position – collision with cargo during device connection – lifting begins – lifting completes – mobile device moves to lowering position – lowering begins – lowering completes – device returns; A matching association is established between the device and the sensing card to identify and monitor cargo information lifted by the monitoring device. This is achieved by making motion time, weight, and position decisions based on data messages uploaded by the sensing card and the weight and position device, and then fusing these three decisions to obtain the decision probability for all cargoes. The cargo with the highest probability is selected as the matching cargo, triggering relevant record management operations for that cargo and alarms for unauthorized lifting. Figure 1 The dashed arrow indicates that this message frame may not necessarily be generated. Please refer to [link / reference]. Figure 2 The overall matching process in the process; S2. Motion Matching Decision: Matching decisions are made using the motion characteristics of the monitoring equipment and the cargo. The motion information of the monitoring equipment is provided by the weight and position device, and the motion information of the cargo is provided by the sensing card. The motion information used for motion matching includes motion direction, motion speed, and motion displacement. The specific implementation manner is calculating the total movement duration reported by each sensing card during the hoisting period, landing period, and cargo-carrying movement period of the supervision equipment. Due to the requirement of the alarm time limit, the movement duration of the supervision equipment during the hoisting period is used as the matching basis. The hoisting period takes the reception of the start hoisting frame as the reference, traces back the L0 period forward and tracks the Li period backward. Assuming that the start hoisting moment is Tr, the total matching period is (Tr-L0, Tr+Li), and it is required that L1<Li<L2. Within this allowable time range, the system automatically performs multiple matching continuously, and the minimum time interval between two adjacent matchings is l, that is, the multiple matching moments are Tr+L1, Tr+L1+l, Tr+L1+2l, ... and so on in sequence, until exceeding Tr+L2, wherein L0 is a parameter representing the length of the retrograde period required for matching, L1 and L2 are the minimum and maximum matchable duration parameters respectively, and l represents the time interval between two consecutive matchings; The matching period duration L involved in movement time matching is defined as (L0+Li), wherein L0 is a parameter representing the length of the period before the start of hoisting that needs to be retrograded for matching; Li is the actual length of time from the start of hoisting when the matching operation is actually performed, satisfying L1<Li<L2, and L1 and L2 are the minimum and maximum matchable durations respectively; Wherein the calculation flow of each movement matching decision is as shown in Figure 3 , Figure 3 wherein C0 is the matching time threshold; S3. Position and weight matching decision: for both position data and weight data, the confidence degrees of position and weight are calculated respectively according to the position or weight matching decision calculation flow chart; Wherein the position or weight matching decision calculation flow chart is as shown in Figure 4 , Figure 4 wherein Ax refers to weight / position data. When Ax refers to weight data, the weight matching decision result is output; when Ax refers to position data, the position matching decision result is obtained, and A0 therein is a matching threshold, which respectively refers to the matching weight threshold W0 and the matching position threshold X0; S4. Calculation of decision fusion strategy probability: according to the movement matching decision calculation flow in step S2 and the position and weight matching decision calculation flow in step S3, the movement matching decision confidence SI, the weight matching decision confidence WI, and the position matching decision confidence XI are calculated respectively, and the decision probabilities of each cargo are respectively , the cargo with the maximum decision probability is selected as the matched cargo, if , it is considered that the matching is unsuccessful; S5. Filtering algorithm: the filtering part adopts a mean-median filtering combined with single iteration prediction filtering algorithm, and the prediction filtering algorithm is as follows: Where R is the algorithm parameter. In mean-median filtering, the number of medians n and the total filter window length N are algorithm parameters. Taking into account the filtering requirements of the complex on-site environment and the limited computing power and storage space of the embedded environment, the filtering part of this algorithm adopts a combination of mean and median filtering and single-iteration prediction filtering algorithm. After testing with actual data, it can be found that the filtering algorithm can effectively filter out on-site noise, while preserving the waveform of the lifting process well. S6. Static / Dynamic Detection Algorithm: A state machine method is used for data smoothing and delay detection. The state machine process is as follows: State 0: if feature_var>=THR1 Accumulator 1 If counter 1 is greater than or equal to CNT1, then switch the state to 1 and clear counter 1. else Clear counter 1 State 1: if feature_var<=THR2 Accumulator 2 If counter 2 is greater than or equal to CNT2, then switch the state to 0 and clear counter 2. else Clear counter 2 In the angular velocity variance window WIN and the delay detection filtering time DELAY, the algorithm parameters are used. In the state machine method, THR1, THR2, CNT1, and CNT2 are all algorithm parameters.

[0013] Actual data testing revealed that the IMU is sensitive to the lifting process. Therefore, the angular velocity variance feature was used to replace the original acceleration variance as the motion feature to be detected by the algorithm. A state machine method was used for data smoothing and delay detection, and the judgment results were time-filtered to remove short-term interference in the actual environment.

[0014] The sensing card is a smart device with a built-in IMU sensor, featuring motion detection and tamper protection functions. The IMU sensor can be simplified to an accelerometer, and its built-in algorithm is a self-developed algorithm. The weight location device refers to a system-integrated device that collects weight location information. Examples of weight devices include overhead crane scales, forklift scales, sensing boxes, and smart pallets. Examples of location devices include cassette positioning systems, laser ranging systems, laser positioning systems, lidar systems, and video AI detection systems.

[0015] Using this algorithm and system, automated monitoring of bulk commodities such as steel coils can be achieved. If steel coils are illegally hoisted without authorization, an audible and visual alarm will be triggered, notifying relevant personnel. Simultaneously, surveillance cameras and gate barriers will be activated to collect evidence and intercept the unauthorized movement, thus effectively ensuring the safety of movable assets. The algorithm design takes into account the numerous challenges the system faces in real-world environments, exhibiting strong applicability, high stability, low power consumption and resource consumption, and strong scalability, allowing for upgrades or customized development for more complex environments or specific scenarios. Using an IMU-based sensing card system to replace traditional methods, the sensing card utilizes sensors such as the IMU to effectively detect the steel coil hoisting process. It can also automatically establish a matching relationship with monitoring equipment—operating machinery pre-installed with relevant equipment, including overhead cranes and forklifts. This enables automatic and effective monitoring of movable assets such as steel coils, providing technological possibilities for movable asset pledging and securitization.

[0016] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A sensor card matching method for warehouse supervision, characterized in that, Comprising the following steps: S1, Supervising the hoisting process of equipment: The complete process of supervising equipment hoisting includes: moving the equipment to the hoisting position—collision with goods during equipment connection—starting hoisting—completing hoisting—moving the equipment to the lowering position—starting lowering—completing lowering—equipment returning; establishing a matching association relationship between the equipment and the perception cards to realize the identification of the information of the goods hoisted by the supervising equipment, and the establishment method is: making motion matching decision, weight matching decision and position matching decision respectively according to the data messages uploaded by the perception cards and the weight-position equipment, fusing the three decisions to obtain the decision probabilities of all goods, selecting the goods with the highest probability as the matching goods, and triggering relevant record management operations for the goods and relevant alarms in case of unauthorized hoisting; S2, The motion matching decision: performing matching decision by using the motion characteristics of the supervising equipment and the goods, the motion characteristics of the supervising equipment are provided by the weight-position equipment, and the motion characteristics of the goods are provided by the perception cards. The motion characteristics used for motion matching include motion direction, motion speed and motion displacement; The specific implementation method is: calculating the total motion duration reported by each perception card during the hoisting period, lowering period and cargo-carrying moving period of the supervising equipment; wherein, due to the requirement of alarm time limit, the motion duration during the hoisting period of the supervising equipment is used as the matching basis; the matching objects are the total motion duration reported by each perception card respectively and the motion duration during the hoisting period of the supervising equipment; taking the time Tr when the start hoisting frame is received as a reference, tracing back L0 duration forward and tracking Li duration backward, the matching period is (Tr-L0, Tr+Li), wherein L0 is the time length parameter that needs to be traced back to before the start of hoisting for matching, Li is the time length parameter from the start of hoisting when the matching operation is actually executed, and L1<Li<L2, L1 is the minimum matchable duration parameter, and L2 is the maximum matchable duration parameter; Within the matching period, the system automatically performs multiple matches continuously, the time interval between two adjacent matches is Δt, that is, the multiple matching moments are Tr+L1, Tr+L1+Δt, Tr+L1+2Δt,... in sequence, until exceeding Tr+L2, wherein Δt is the time interval parameter between two consecutive matches; The total duration L of the matching period for motion matching is L=L0+Li; in each matching, traversing the historical motion data of the perception cards, calculating the cumulative motion duration of the perception card within the matching period as the motion matching length Sc, and comparing Sc with a preset matching time threshold C0: When Sc≤C0, the confidence SI of the motion matching decision for this matching is 0; When Sc>C0, the confidence SI of the motion matching decision for this matching is (Sc-C0) / (L-C0); S3, The position and weight matching decision: position confidence or weight confidence is calculated by position or weight matching decision methods respectively for position or weight data; wherein the position or weight matching decision comprises: Obtaining the measured value Ai of position or weight, traversing the recorded value Ab of each goods, calculating the difference degree Ac=|Ai-Ab| between the measured value and the recorded value, and comparing Ac with a preset matching threshold A0: When Ac > A0, the corresponding matching decision confidence is 0; When Ac≤A0, the corresponding matching decision confidence is min(1, 1.1-Ac / A0). When Ac is location data, the output is the location matching decision confidence score XI; when Ac is weight data, the output is the weight matching decision confidence score WI. S4. Calculation of the probability of the decision fusion strategy: Based on the motion matching decision calculation process in step S2 and the position and weight matching decision calculation process in step S3, the motion matching decision confidence S1, weight matching decision confidence S1, and position matching decision confidence S1 are calculated respectively. A fusion calculation is performed based on the motion matching decision confidence S1, weight matching decision confidence S1, and position matching decision confidence S1 for each cargo to obtain the decision probability Pj for each cargo. The cargo with the highest decision probability Pj is selected as the matching cargo. The fusion calculation method for the decision probability Pj is as follows: ;like If the match is not found, then the match is considered unsuccessful.

2. The sensor card matching method for warehouse supervision according to claim 1, characterized in that: The sensing card has a built-in IMU sensor.

3. The sensor card matching method for warehouse supervision according to claim 1, characterized in that: The weight location device is a system-integrated device used to collect weight and location information.

Images