Automatic pick-up management method and system based on low-altitude express cabinet

Abstract

The invention discloses an automatic pick-up management method and system based on a low-altitude express cabinet, and relates to the related field of logistics management, and the method comprises the steps: building an in-cabinet process mechanism model based on structure parameter priori knowledge; dynamic sensing data in the express delivery cabinet entering process is collected through a multi-mode sensing unit; performing physical logic consistency verification on the dynamic sensing data by using the in-cabinet process mechanism model to generate pickup management information; constructing a cabinet body state mechanism model of the target cabinet grid based on the pick-up management information in combination with the mechanical characteristic parameters of the low-altitude express cabinet; after the pick-up task is triggered, generating a pick-up instruction based on the cabinet state mechanism model; and the pick-up instruction and the one-time pick-up voucher are issued to the logistics robot, and the logistics robot is driven to execute automatic cabinet opening pick-up operation. The technical problem that existing automatic pick-up management is insufficient in overall safety controllability is solved, and the technical effect of improving the safety controllability of the whole process of automatic pick-up of the low-altitude express cabinet is achieved.

Description

Technical Field

[0001] This application relates to the field of logistics management, and in particular to an automatic package pickup management method and system based on low-altitude express cabinets. Background Technology

[0002] Automated parcel pickup management at the last mile of low-altitude unmanned delivery is a crucial link in ensuring delivery safety, improving logistics efficiency, and achieving large-scale commercialization, significantly impacting the reliability of the entire low-altitude logistics system. Currently, the industry generally employs traditional fixed-process control for managing smart parcel locker terminals, relying primarily on single switch signals, weight detection, or manual confirmation to complete the locker entry and retrieval actions, executing locker door opening and closing and equipment scheduling according to preset programs. This traditional approach relies solely on simple sensors and fixed logic, lacking in-depth analysis of dynamic interactions. This leads to inaccurate determination of parcel safety upon entry, long-term susceptibility to abnormal impacts causing structural hazards in the locker, and a mismatch between automated retrieval actions and locker status, easily resulting in collisions, jamming, or structural fatigue, making it difficult to guarantee overall stability and safety.

[0003] At present, the automatic package pickup management of low-altitude express lockers has the technical problem of insufficient overall safety and controllability. Summary of the Invention

[0004] This application provides an automated parcel retrieval management method and system based on low-altitude parcel lockers. It employs a physical process model of parcel entry into the locker based on the hardware structure of the locker, collects multi-dimensional real-time data during parcel delivery, uses the physical model to verify compliance of the entry process, establishes a state model for the corresponding locker compartment based on the verification results, and generates a safe retrieval instruction based on this model during retrieval. This instruction, along with a credential, is then sent to a robot to complete the automated retrieval. These technical means solve the technical problem of insufficient overall safety and controllability in existing automated parcel retrieval management systems for low-altitude parcel lockers, achieving a significant improvement in the overall safety and controllability of the automated parcel retrieval process.

[0005] This application provides an automatic package retrieval management method based on low-altitude express lockers, comprising: constructing a locker entry process mechanism model based on prior knowledge of the structural parameters of the low-altitude express locker; when delivering packages at low-altitude delivery terminals, collecting dynamic sensing data of the package entry process through the multimodal sensing unit built into the low-altitude express locker; using the locker entry process mechanism model to perform physical and logical consistency verification on the dynamic sensing data, generating package retrieval management information; based on the package retrieval management information and combined with the mechanical characteristic parameters of the low-altitude express locker, constructing a locker state mechanism model of the target locker compartment; after the retrieval task is triggered, generating a retrieval instruction based on the locker state mechanism model; and sending the retrieval instruction and a one-time retrieval voucher to a logistics robot to drive the automatic locker opening and retrieval operation.

[0006] In a possible implementation, based on prior knowledge of the structural parameters of the low-altitude express delivery locker, a mechanism model of the locker entry process is constructed, and the following processing is performed: obtaining the cabinet compartment geometry, material properties, and sensor placement parameters of the low-altitude express delivery locker to obtain the prior knowledge of the structural parameters; and establishing the mechanism model of the locker entry process based on the prior knowledge of the structural parameters.

[0007] In a possible implementation, when a parcel is delivered at a low-altitude delivery terminal, dynamic sensing data during the parcel's entry into the locker is collected by a multimodal sensing unit built into the low-altitude parcel locker, and the following processing is performed: pressure timing signals during the parcel's entry into the locker are collected by a distributed pressure sensing unit; trajectory and posture image sequences during the parcel's entry into the locker are collected by a visual sensing unit; and micro-strain signals of the locker structure during the parcel's entry into the locker are collected by a strain sensing unit. After data cleaning of the pressure timing signals, the trajectory and posture image sequences, and the micro-strain signals of the locker structure, the dynamic sensing data is obtained.

[0008] In a possible implementation, the physical and logical consistency of the dynamic sensing data is verified using the cabinet entry process mechanism model to generate package retrieval management information, and the following processing is performed: the instantaneous measured velocity of the package upon cabinet entry collision is calculated from the trajectory posture image sequence in the dynamic sensing data; the measured mass of the package is calculated from the pressure time series signal in the dynamic sensing data; the instantaneous measured velocity upon cabinet entry collision and the measured mass of the package are input into the cabinet entry process mechanism model to calculate the theoretical impact force distribution and theoretical strain response; the measured impact force distribution is obtained by spatiotemporal integration inversion of the pressure time series signal; the theoretical impact force distribution and the measured impact force distribution are compared for the first time; the theoretical strain response is compared with the cabinet structure micro-strain signal in the dynamic sensing data for the second time; when the deviations of the first comparison and the second comparison are both within the preset physical law tolerance range, the verification is deemed successful, and the package retrieval management information is generated accordingly.

[0009] In a possible implementation, based on the package pickup management information and combined with the mechanical characteristic parameters of the low-altitude express cabinet, a cabinet state mechanism model of the target cabinet is constructed, and the following processing is performed: extract the physical parameters of the package, the cabinet location information, and the initial stress state after entering the cabinet from the package pickup management information to generate the initial conditions of the model; based on the initial conditions of the model, establish the cabinet state mechanism model.

[0010] In a possible implementation, after the retrieval task is triggered, a retrieval command is generated based on the cabinet state mechanism model, and the following processing is performed: based on the retrieval task issued by the background management system, the predetermined retrieval time window and environmental wind speed factors are input as boundary conditions into the cabinet state mechanism model for simulation calculation; the simulation results are obtained, including the optimal opening speed curve and angle curve of the cabinet door that meet the mechanism safety constraints, and the standard docking posture of the robotic arm; the optimal opening speed curve of the cabinet door, the angle curve, and the standard docking posture of the robotic arm are encapsulated to generate the retrieval command.

[0011] In a possible implementation, the pickup instruction and one-time pickup voucher are sent to a logistics robot to drive an automatic locker opening and pickup operation. The following processes are performed: the backend management system binds the pickup instruction and one-time pickup voucher and sends them to the assigned logistics robot; the logistics robot navigates to the low-altitude express locker and sends the one-time pickup voucher to the locker for identity authentication; after successful authentication, the logistics robot adjusts its end effector according to the standard docking posture of the robotic arm in the pickup instruction and physically docks with the locker door mechanism, triggering an opening signal; the low-altitude express locker, based on the locker state mechanism model, verifies that the current state meets the mechanism safety constraints in the pickup instruction and then drives the locker door locking mechanism to unlock; the logistics robot, based on the optimal opening speed curve and angle curve of the locker door in the pickup instruction, controls the locker door to open and retrieve the package.

[0012] In a possible implementation, after the pickup instruction and one-time pickup voucher are issued to the logistics robot to drive the automatic cabinet opening and pickup operation, the following processing is also performed: After the logistics robot controls the cabinet door to close, the multimodal sensing unit collects the full-cycle actual dynamic data of the cabinet door closing process; the full-cycle actual dynamic data is compared in real time with the predicted dynamic data of the cabinet state mechanism model for the same closing process; if the comparison deviation is within the preset safety tolerance range, the pickup operation is deemed compliant, and the prediction accuracy of the cabinet state mechanism model under the current working condition is confirmed; if the comparison deviation exceeds the safety tolerance range, the model prediction is deemed inaccurate, and the cabinet state mechanism model is reverse-calibrated and optimized based on the full-cycle actual dynamic data.

[0013] In possible implementations, the following processing is also performed: The backend management system calls the cabinet state mechanism model, performs simulation prediction to obtain scheduling decision basis, and performs at least one of the following scheduling operations: marking parcels whose displacement risk value predicted by the cabinet state mechanism model is higher than the threshold as high priority, and assigning priority retrieval tasks to the logistics robots; when a new parcel needs to be stored, querying the structural fatigue value predicted by the cabinet state mechanism model for each empty cabinet, and prioritizing the allocation of the new parcel to the cabinet with the lowest current fatigue value; when multiple logistics robots need to retrieve parcels from different cabinets of the same low-altitude express cabinet, simulating the overall vibration response of the cabinet under different operation sequences of each robot based on the cabinet state mechanism model, calculating a multi-task execution time sequence to avoid overall vibration superposition or resonance, and controlling the robots to perform retrieval operations sequentially according to the multi-task execution time sequence.

[0014] This application also provides an automated parcel pickup management system based on low-altitude express lockers, comprising: a locker entry process mechanism model construction module, used to construct a locker entry process mechanism model based on prior knowledge of the structural parameters of the low-altitude express locker; a dynamic sensing data acquisition module, used to collect dynamic sensing data of the parcel entry process through the multimodal sensing unit built into the low-altitude express locker when the parcel is delivered at the low-altitude delivery terminal; a physical logic consistency verification module, used to perform physical logic consistency verification on the dynamic sensing data using the locker entry process mechanism model, and generate pickup management information; a locker state mechanism model construction module, used to construct a locker state mechanism model of the target locker based on the pickup management information and the mechanical characteristic parameters of the low-altitude express locker; a pickup instruction generation module, used to generate a pickup instruction based on the locker state mechanism model after the pickup task is triggered; and an automatic locker opening and pickup module, used to send the pickup instruction and a one-time pickup voucher to the logistics robot to drive the automatic locker opening and pickup operation.

[0015] This application proposes an automated parcel retrieval management method and system based on low-altitude express lockers. First, based on prior knowledge of the structural parameters of the low-altitude express locker, a mechanism model of the parcel entry process is constructed. Then, when a parcel is delivered from a low-altitude delivery terminal, dynamic sensing data during the parcel entry process is collected by the multimodal sensing unit built into the low-altitude express locker. Next, the mechanism model of the entry process is used to verify the physical and logical consistency of the dynamic sensing data, generating retrieval management information. Based on this retrieval management information and the mechanical characteristic parameters of the low-altitude express locker, a locker state mechanism model of the target locker is constructed. When a retrieval task is triggered, a retrieval command is generated based on the locker state mechanism model. Finally, the retrieval command and a one-time retrieval voucher are sent to a logistics robot to drive the automatic locker opening and retrieval operation. Through the above process, the method and system proposed in this application achieve the technical effect of improving the safety and controllability of the entire automated parcel retrieval process of low-altitude express lockers. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating an automatic package pickup management method based on low-altitude express lockers, provided as an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a low-altitude express cabinet automatic package pickup management system provided in an embodiment of this application.

[0019] Explanation of reference numerals in the attached diagram: Module 10 for constructing the mechanism model of the cabinet entry process, Module 20 for dynamic sensing data acquisition, Module 30 for physical and logical consistency verification, Module 40 for constructing the mechanism model of the cabinet status, Module 50 for generating the retrieval instruction, and Module 60 for automatically opening the cabinet and retrieving items. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structure, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides an automatic package pickup management method based on low-altitude express lockers, such as... Figure 1 As shown, the method includes: Step S100: Based on prior knowledge of the structural parameters of the low-altitude express cabinet, construct a mechanism model of the cabinet entry process.

[0022] Specifically, based on the pre-determined hardware structure information of the low-altitude express delivery locker, a mathematical and physical model is established to accurately describe the physical changes of the entire process of placing a package into the locker compartment, used to verify whether the placement is safe and compliant. Specifically, a proportional geometric model of the low-altitude express delivery locker is established using 3D modeling software. Then, based on the finite element analysis method, the locker body and compartments are divided into mesh elements. The elastic modulus, Poisson's ratio, and density parameters of the locker material, as well as the length, width, depth, and wall thickness parameters of the compartments, are input. Sensor installation locations are used as model observation nodes to construct a mechanism model of the placement process that includes kinematic and dynamic equations.

[0023] In one possible implementation, based on prior knowledge of the structural parameters of the low-altitude express delivery locker, a mechanism model of the locker entry process is constructed. Step S100 further includes step S110, acquiring the geometric dimensions, material properties, and sensor placement parameters of the locker compartments to obtain the prior knowledge of the structural parameters. Specifically, the basic hardware parameters of the low-altitude express delivery locker are collected and organized to provide raw data support for constructing the locker entry process mechanism model. Specifically, the length, width, and depth dimensions inside the locker compartments are measured point by point using a laser rangefinder. The yield strength, elastic modulus, and density parameters of the locker material are tested using a material mechanics testing machine. The three-dimensional coordinates of each pressure sensor, vision sensor, and strain sensor are located using a total station, and their positional relationship relative to the locker compartment is recorded. For example, the measured dimensions of the locker compartment are 40 cm long, 30 cm wide, and 20 cm high. The locker material is aluminum alloy with a density of 2.7 g / cm³. The pressure sensors are placed at the four corners of the bottom of the locker compartment, with coordinates of 5 cm from the left side and 5 cm from the front side, etc., forming a structured parameter document.

[0024] Step S120: Based on the prior knowledge of the structural parameters, establish a mechanism model for the package insertion process. Specifically, the collected prior knowledge of the low-altitude express locker structural parameters is used as the core input to build a mechanism model that can accurately characterize the physical process of package insertion into the locker. This model can fully reflect the kinematics, dynamics, and locker structural response laws when packages are inserted into the locker. Specifically, a three-dimensional explicit dynamic model was constructed based on the finite element method, and the Lagrange algorithm was selected as the core solution algorithm to construct the core dynamic equations of the cabinet entry process. The impact force was calculated by the difference in velocity before and after the collision and the collision contact time, and the cabinet strain was calculated by the ratio of stress to the elastic modulus of the cabinet material. The cabinet geometry, material properties, and sensor position parameters were mapped to the model's mesh elements, material parameters, and observation nodes, respectively. The mesh was divided and the material model and contact type were defined. The contact stiffness coefficient and damping coefficient were set, and the sensor positions were set as the model node set to output the impact force and strain data at the corresponding positions. Then, the implicit parameters such as the model damping coefficient and contact stiffness were calibrated by small sample measured data to ensure that the model error was controlled within 5%, thus completing the establishment and verification of the cabinet entry process mechanism model.

[0025] Step S200: When delivering a parcel at a low-altitude delivery terminal, dynamic sensing data during the parcel's entry into the parcel locker is collected through the multimodal sensing unit built into the low-altitude parcel locker.

[0026] Specifically, when a drone or low-altitude delivery robot places a package into the locker, multiple types of sensors simultaneously collect physical signals during the placement process, generating multi-dimensional dynamic data. Specifically, the distributed pressure sensing unit, visual sensing unit, and strain sensing unit built into the low-altitude locker are activated, and a synchronized acquisition clock is set. Signals from the package's entry into the locker area to its stable placement are collected at the same sampling frequency, completing the initial data acquisition. For example, the three sensor units can be set to a synchronized sampling frequency of 100 Hz, triggering acquisition from the moment the package enters the locker area until it comes to a complete stop.

[0027] In one possible implementation, when a parcel is delivered at a low-altitude delivery terminal, dynamic sensing data during the parcel's entry into the locker is collected by a multimodal sensing unit built into the locker. Step S200 further includes step S210, where a distributed pressure sensing unit collects pressure timing signals during the parcel's entry into the locker. Specifically, multiple pressure sensors deployed at the bottom of the locker continuously collect pressure change data at different locations during the parcel's entry. Specifically, five thin-film pressure sensors are installed at the four corners and center of the locker bottom. The sensor range is set to 0 to 50 kg, and the sampling rate is set to 100 Hz. When the parcel is entered, the sensors output voltage signals in real time. The analog signals are converted into digital pressure values ​​by an analog-to-digital converter, forming a pressure timing signal arranged in chronological order.

[0028] Step S220: The visual sensing unit acquires a sequence of trajectory and posture images during the process of the package being placed in the locker. Specifically, a camera inside the locker continuously captures images of the entire process of the package being placed in the locker, recording the package's movement trajectory and posture changes. Specifically, a binocular depth camera is installed at the locker entrance, with a resolution set to 1280×720. Images are continuously captured starting from when the package enters the shooting area. Each frame records the package's position, tilt angle, and rotation state, forming a time-ordered image sequence. For example, 30 frames are continuously captured, recording the complete trajectory and posture of the package as it falls from a 45-degree angle to a horizontal position.

[0029] Step S230: The micro-strain signals of the cabinet structure during the package insertion process are collected using a strain sensing unit. Specifically, strain gauges attached to the load-bearing parts of the cabinet are used to collect the micro-deformation signals of the cabinet caused by the impact of the package entering the cabinet. Specifically, resistive strain gauges are attached to the load-bearing beams on the side walls and bottom of the cabinet compartment. The strain gauge sensitivity coefficient is set to 2.0. The deformation is converted into a voltage signal using a Wheatstone bridge, and the sampling rate is set to 100 Hz. The micro-strain values ​​generated by the impact of the package are collected in real time. For example, when the package is inserted into the cabinet, the strain gauge at the bottom of the compartment collects a micro-strain rising from 0 to 80 micro-strain, forming a continuous strain time-series signal.

[0030] Step S240 involves cleaning the pressure time-series signal, the trajectory attitude image sequence, and the cabinet structure micro-strain signal to obtain the dynamic sensing data. Specifically, noise, outliers, and duplicate data are removed from the acquired signals to obtain clean and usable dynamic sensing data. Specifically, a sliding window mean filter is used to remove high-frequency noise from the pressure and strain signals, with a window length of 10 sampling points. Outliers exceeding the sensor's range are eliminated. The image sequence is deduplicated and deblurred, retaining clear and valid frames. Simultaneously, the time axes of the three signals are aligned; for example, outliers greater than 50 kg in the pressure signal are replaced with the mean of adjacent data, and blurred image frames are deleted to ensure a one-to-one correspondence between the pressure, image, and strain data at the same moment.

[0031] Step S300: Use the cabinet entry process mechanism model to perform physical and logical consistency verification on the dynamic sensing data and generate item retrieval management information.

[0032] Specifically, the actual collected data is compared with the theoretical calculation results of the locker entry process mechanism model to determine whether the locker entry process conforms to physical laws, generating basic information for subsequent package retrieval management. Specifically, the cleaned dynamic sensing data is input into the locker entry process mechanism model to calculate the theoretical physical response. Deviation is calculated between this model and the measured data to determine if the deviation is within the allowable range. If the verification passes, package retrieval management information including package parameters and locker status is generated. For example, the measured speed and weight are input into the locker entry process mechanism model to obtain the theoretical impact force and strain, which are then compared with the measured values ​​to determine compliance and generate corresponding management information.

[0033] In one possible implementation, the physical and logical consistency of the dynamic sensing data is verified using the cabinet entry process mechanism model to generate package retrieval management information. Step S300 further includes step S310, which involves parsing and calculating the instantaneous measured velocity of the package upon collision with the cabinet from the trajectory posture image sequence in the dynamic sensing data. Specifically, the actual velocity of the package at the instant of contact with the cabinet is calculated using the cabinet entry image sequence. Specifically, target detection and tracking are performed on the image sequence acquired by a binocular depth camera to determine the three-dimensional spatial coordinates of the package in each frame. The instantaneous velocity is calculated based on the time interval and displacement difference between adjacent frames, and the velocity of the package in the frame before contact with the cabinet is extracted as the instantaneous measured velocity upon collision.

[0034] Step S320: The measured mass of the package is calculated from the pressure time-series signal in the dynamic sensing data. Specifically, the actual mass of the package is calculated using the stable pressure value at the bottom of the compartment. Specifically, the stable pressure value of the package after it comes to rest is extracted from the pressure time-series signal, the pressure values ​​of multiple bottom sensors are summed, and the measured mass of the package is calculated using the gravity conversion formula: mass equals total pressure divided by gravitational acceleration. For example, if the sum of the stable pressures from five sensors is 147 Newtons, and the gravitational acceleration is taken as 9.8 m / s², the calculated mass of the package is 15 kg.

[0035] Step S330: The instantaneous measured velocity of the collision upon entering the cabinet and the measured mass of the package are input into the cabinet entry process mechanism model to calculate the theoretical impact force distribution and theoretical strain response. Specifically, the measured velocity and mass are substituted into the cabinet entry process mechanism model to calculate the theoretically expected impact force and cabinet strain. Specifically, the instantaneous measured velocity of the collision upon entering the cabinet and the measured mass of the package are used as input parameters, imported into the cabinet entry process mechanism model, and a dynamic simulation calculation is run to output the theoretical impact force values ​​at each point on the bottom of the cabinet and the theoretical micro-strain values ​​at each strain measurement point on the cabinet.

[0036] Step S340: The measured impact force distribution is obtained by performing spatiotemporal integration inversion on the pressure time-series signal, and a first comparison is made between the theoretical impact force distribution and the measured impact force distribution. Specifically, the discrete pressure time-series signal is converted into impact force distribution data in the cabinet space dimension using mathematical methods, and then quantitatively compared with the theoretical impact force distribution calculated by the cabinet entry process mechanism model to verify the physical consistency of the cabinet entry process. Specifically, the time integration window is determined to be from the collision initiation time to the pressure stabilization time. The pressure time-series signal of each sensor is integrated in the time domain to obtain the impulse at each sensor position. Then, Kriging interpolation is used to extend the impulse distribution to the entire bottom surface of the cabinet with the same resolution as the cabinet grid unit. Combined with the collision contact time inversion, the measured impact force of each grid is obtained to form a complete measured impact force distribution. The theoretical impact force distribution and the measured impact force distribution output by the cabinet entry process mechanism model are mapped to a unified cabinet grid coordinate system. The absolute deviation, global root mean square error, and maximum deviation value of each grid are calculated. The global root mean square error is obtained by taking the square root of the sum of the squares of the differences between the measured and theoretical impact forces of each grid divided by the total number of grids. Combined with the preset root mean square error and maximum deviation tolerance threshold, the first comparison between the theoretical impact force distribution and the measured impact force distribution is completed, and it is determined whether the comparison result meets the requirements.

[0037] Step S350: Perform a second comparison between the theoretical strain response and the cabinet structure micro-strain signal in the dynamic sensing data. Specifically, compare the theoretical strain calculated by the cabinet insertion process mechanism model with the actual collected cabinet strain. Specifically, extract the measured micro-strain peak value at the moment of collision from the cabinet structure micro-strain signal, calculate the difference with the corresponding theoretical micro-strain value output by the cabinet insertion process mechanism model, obtain the strain deviation value, and complete the second comparison.

[0038] Step S360: When the deviations of the first comparison and the second comparison are both within the preset physical law tolerance range, the verification is deemed successful, and the package retrieval management information is generated accordingly. Specifically, when both the impact force deviation and the strain deviation are within the safe allowable range, the cabinet entry is confirmed to be compliant, and the management information required for package retrieval is generated. Specifically, the preset impact force tolerance range is ±10 Newtons, and the strain tolerance range is ±10 microstrains. When both the impact force deviation and the strain deviation are within the corresponding ranges, the verification is deemed successful, and the measured quality of the package, the collision speed, the cabinet number, the initial stress state of the cabinet, etc., are packaged into the package retrieval management information.

[0039] Step S400: Based on the package pickup management information and combined with the mechanical characteristic parameters of the low-altitude express cabinet, construct a cabinet state mechanism model of the target cabinet.

[0040] Specifically, based on the locker entry verification results and the locker's mechanical parameters, a dedicated state model is established for the target locker where the parcel is stored. Specifically, parcel parameters and locker information are extracted from the parcel retrieval management information. Combined with the locker's stiffness, damping, and natural frequency mechanical characteristics, a locker state mechanism model is established on top of the existing locker entry process mechanism model to describe the real-time state of the locker, simulating locker vibration and stress changes during parcel retrieval.

[0041] In one possible implementation, based on the package pickup management information and combined with the mechanical characteristic parameters of the low-altitude express cabinet, a cabinet state mechanism model of the target cabinet is constructed. Step S400 further includes step S410, which extracts the physical parameters of the package, the cabinet location information, and the initial stress state after the package is placed in the cabinet from the package pickup management information to generate the initial conditions for the model. Specifically, the initial data required for modeling is extracted from the package pickup management information to set the initial state of the model. Specifically, the package pickup management information data packet is parsed to extract the physical parameters of the package's mass, size, and center of gravity position, the cabinet number, spatial coordinate position information, and the initial stress and initial deformation state of the cabinet after it is placed in the cabinet. The above data is used as the initial input conditions for the model.

[0042] Step S420: Based on the initial conditions of the model, establish the cabinet state mechanism model. Specifically, based on the extracted physical parameters of the parcel, cabinet location information, and initial stress state, and combined with the mechanical property parameters of the cabinet, construct a mechanism model that can characterize the structural state of the target cabinet in real time, such as stress, strain, vibration, and fatigue, providing a core basis for generating parcel retrieval instructions. Specifically, using the core logic of combining initial condition assignment, mechanical property coupling, and state equation construction, the parcel mass in the initial conditions of the model is first converted into a uniformly distributed load, and the initial stress is used as a preload, which are mapped to the initial boundary conditions of the model, located to the finite element mesh of the target cabinet, and the initial condition loading is completed. Then, the stiffness matrix, damping matrix, and natural frequency in the mechanical property parameters of the cabinet are introduced to construct a multi-degree-of-freedom vibration equation, wherein the damping matrix is ​​calculated through the damping ratio, mass matrix, and stiffness matrix, and the equation includes the correlation between the mass matrix, damping matrix, stiffness matrix, and acceleration, velocity, displacement, and external load. A fatigue accumulation calculation module was introduced, and based on the Miner linear fatigue accumulation criterion, the cabinet strain time series data was converted into fatigue degree, constructing a multi-dimensional cabinet state mechanism model that includes static stress, dynamic vibration, and fatigue accumulation. Finally, modal analysis was used to verify whether the deviation between the model's natural frequency and the measured cabinet's natural frequency was controlled within 5%, completing model calibration and ensuring that the model can accurately output the real-time structural state parameters of the target cabinet.

[0043] Step S500: After the item retrieval task is triggered, an item retrieval instruction is generated based on the cabinet state mechanism model.

[0044] Specifically, upon receiving a pickup request, the system calculates the control parameters for safe pickup using the cabinet state mechanism model, and generates standardized pickup instructions. Specifically, after the backend management system issues a pickup task, it calls the cabinet state mechanism model for real-time simulation, calculates the cabinet door motion parameters and robot docking parameters that satisfy safety constraints, and encapsulates them into directly executable pickup instructions. For example, the simulation obtains the cabinet door opening speed, angle, and robot pose, generating instructions containing specific values.

[0045] In one possible implementation, after the pickup task is triggered, a pickup instruction is generated based on the cabinet state mechanism model. Step S500 further includes step S510, which, based on the pickup task issued by the background management system, inputs the predetermined pickup time window and environmental wind speed as boundary conditions into the cabinet state mechanism model for simulation calculation. Specifically, the pickup time and external environmental conditions are input into the model to simulate the cabinet state under real working conditions. Specifically, the predetermined pickup start time and duration are extracted from the pickup task to form a time window, the current environmental wind speed is obtained through a meteorological sensor, and the time window, wind speed, and initial cabinet state are input into the model as boundary conditions to run transient dynamics simulation.

[0046] Step S520: Obtain simulation results, including the optimal opening speed curve and angle curve of the cabinet door that satisfy the mechanistic safety constraints, as well as the standard docking posture of the robotic arm. Specifically, extract safe and stable cabinet door motion parameters and robot docking posture from the simulation results. Specifically, screen for safe operating conditions in the simulation results where the cabinet stress is less than the yield strength and the vibration amplitude is less than 2 mm; extract the speed-time curve and angle-time curve of the cabinet door from fully closed to fully open, as well as the three-dimensional coordinates and posture angle of the robot docking with the cabinet door.

[0047] Step S530: Encapsulate the optimal opening speed curve of the cabinet door, the angle curve, and the standard docking pose of the robotic arm to generate the retrieval command. Specifically, the optimal control parameters obtained from simulation are packaged into a standard command format. Specifically, according to a preset communication protocol, the cabinet door speed curve, angle curve, and robotic arm pose data are converted into binary data frames, check bits and cabinet numbers are added, and a retrieval command that meets the robot control requirements is generated.

[0048] In step S600, the pickup instruction and one-time pickup voucher are sent to the logistics robot to drive the automatic cabinet opening and pickup operation.

[0049] Specifically, the backend system sends the pickup instruction and security credentials to the designated robot, controlling the robot to complete the pickup. Specifically, the backend management system sends the pickup instruction and one-time encrypted credentials to the target logistics robot via a 5G or Wi-Fi communication module. After receiving the data, the robot parses it and prepares to execute it. For example, it sends the instruction and credentials to the robot's IP address via the UDP communication protocol, and the robot receives and stores them in its control cache.

[0050] In one possible implementation, the pickup instruction and one-time pickup voucher are sent to the logistics robot to drive the automatic cabinet opening and pickup operation. Step S600 further includes step S610, whereby the back-end management system binds the pickup instruction and one-time pickup voucher and sends them to the assigned logistics robot. Specifically, the instruction and voucher are bound one-to-one and sent to the robot responsible for pickup. Specifically, the back-end management system generates a one-time pickup voucher encrypted with a random number, associates and binds it with the corresponding pickup instruction, and sends it to the designated robot terminal according to the task allocation result. For example, a 16-bit random string is generated as a one-time voucher, bound to the pickup instruction for cabinet 05, and sent to the robot numbered Robot-08.

[0051] In step S620, the logistics robot navigates to the low-altitude express locker and sends the one-time pickup voucher to the locker for identity authentication. Specifically, the robot autonomously drives to the locker and sends the voucher to verify its own permissions. Specifically, the logistics robot, based on LiDAR and visual navigation, travels along a preset path to the docking position of the low-altitude express locker and sends the one-time pickup voucher to the locker control board via a serial communication module. The locker control board compares and verifies the received voucher with the voucher stored in the background.

[0052] Step S630: After successful authentication, the logistics robot adjusts its end effector according to the standard docking posture of the robotic arm in the pickup instruction and physically docks with the cabinet door mechanism, triggering an opening signal. Specifically, after successful authentication, the robot adjusts its robotic arm to the designated position and completes docking with the cabinet door. Specifically, the logistics robot drives the joint motors to adjust the end effector position according to the three-dimensional coordinates and attitude angles in the instruction, locks it with the cabinet door docking structure via a positioning pin, and triggers the Hall sensor to generate an opening signal.

[0053] In step S640, the low-altitude express delivery locker, based on the real-time feedback of the locker's state mechanism model, verifies that the current state meets the mechanism safety constraints in the retrieval instruction, and then drives the lock door locking mechanism to unlock. Specifically, when the real-time state of the locker meets the safety requirements, the lock is controlled to unlock. Specifically, the locker state mechanism model calculates the current stress, vibration, and deformation states in real time, compares them with the safety thresholds in the retrieval instruction, and when the safety requirements are met, the control board drives the electromagnetic lock to perform the unlocking action.

[0054] In step S650, the logistics robot controls the cabinet door to open and retrieve the package according to the optimal opening speed curve and angle curve in the retrieval instruction. Specifically, the robot controls the cabinet door to open smoothly according to the instruction, completing the package retrieval. Specifically, the robot controls the motor to output torque according to the speed curve, causing the cabinet door to open according to the optimal opening speed and angle curves. After opening to the correct position, the robotic arm extends into the cabinet to grip the package, completing the retrieval.

[0055] In one possible implementation, after the pickup instruction and one-time pickup voucher are sent to the logistics robot to drive the automatic cabinet opening and pickup operation, the process further includes step S700: after the logistics robot controls the cabinet door to close, the multimodal sensing unit collects the actual dynamic data of the entire cabinet door closing process. Specifically, when the cabinet door closes after pickup, the dynamic signals of the entire cabinet door closing process are collected again. Specifically, during the process of the logistics robot controlling the cabinet door to close, pressure, vision, and strain sensors are simultaneously activated to collect the full-cycle data of the door's movement speed, impact force, and cabinet deformation when the door closes.

[0056] Step S800 involves comparing the actual dynamic data of the entire cycle with the predicted dynamic data of the cabinet state mechanism model for the same closing process in real time. Specifically, the measured data of the cabinet door closing is compared with the model's predicted data. Specifically, the cabinet state mechanism model predicts the velocity, impact force, and strain data at each moment in advance based on the cabinet door closing command, and calculates the difference between these values ​​and the measured data at the same moment, completing the real-time comparison.

[0057] Step S900: If the comparison deviation is within the preset safety tolerance range, the retrieval operation is deemed compliant, and the accuracy of the cabinet state mechanism model's prediction under the current operating conditions is confirmed. Specifically, if the deviation is within the allowable range, the retrieval is valid, and the model is accurate. Specifically, preset tolerances are set for cabinet door movement speed, impact force, and cabinet deformation; if all deviations are within the specified ranges, the retrieval is deemed compliant, and the model is marked as accurate under the current wind speed and load conditions.

[0058] Step S1000: If the comparison deviation exceeds the safety tolerance range, it is determined that the model prediction is inaccurate. Based on the full-cycle actual dynamic data, the cabinet state mechanism model is reverse-calibrated and optimized. Specifically, when the deviation exceeds the tolerance range, the model parameters are corrected to improve the subsequent prediction accuracy. Specifically, when the deviation exceeds the safety tolerance range, the full-cycle actual dynamic data is extracted, and the stiffness and damping parameters in the model are fitted and corrected using the least squares method. The material properties and boundary conditions of the finite element model are updated to complete the model calibration.

[0059] In one possible implementation, the system further includes step S1100: the backend management system calls the cabinet state mechanism model, performs simulation prediction to obtain scheduling decision basis, and performs at least one of the following scheduling operations: marking parcels whose displacement risk value predicted by the cabinet state mechanism model is higher than a threshold as high priority, and assigning priority retrieval tasks to the logistics robots; when a new parcel needs to be stored, querying the structural fatigue value predicted by the cabinet state mechanism model for each empty cabinet, and prioritizing the allocation of the new parcel to the cabinet with the lowest current fatigue value; when multiple logistics robots need to retrieve parcels from different cabinets of the same low-altitude express cabinet, simulating the overall vibration response of the cabinet under different operation sequences of each robot based on the cabinet state mechanism model, calculating the multi-task execution time sequence to avoid overall vibration superposition or resonance, and controlling the robots to perform retrieval operations sequentially according to the multi-task execution time sequence. Specifically, the backend management system obtains key data such as parcel displacement risk, cabinet structural fatigue value, and multi-robot operation vibration response through the simulation prediction function of the cabinet state mechanism model, and optimizes the express cabinet scheduling strategy based on this to ensure cabinet safety and retrieval efficiency. Specifically, corresponding implementation logic is adopted for different scheduling scenarios. During high-priority pickup scheduling, the physical parameters of all packages currently in the locker and the initial state of the locker are input into the locker state mechanism model. The model simulates and predicts the displacement of each package due to environmental vibration within a preset time period, calculating the displacement risk value. The displacement risk value is the ratio of the predicted displacement to the safe displacement threshold. Packages with risk values ​​exceeding the preset threshold are marked as high-priority, and priority pickup tasks are assigned to the nearest idle logistics robot using a greedy algorithm. During locker allocation scheduling, for all idle lockers, the locker state mechanism model is called to simulate and calculate the current structural fatigue of each locker. The fatigue value is calculated based on the Miner linear fatigue accumulation criterion, and is the ratio of cumulative damage to allowable damage. Packages are sorted from low to high fatigue value, and new packages are assigned to the locker with the lowest fatigue value. If multiple... If all compartments have the same fatigue level, the compartment with the highest spatial matching degree is selected first. The spatial matching degree is calculated as the ratio of the difference between the compartment volume and the parcel volume to the compartment volume. During multi-robot time-series scheduling, the parcel retrieval operation parameters of each robot are input into the cabinet state mechanism model to simulate the overall vibration response of the cabinet under different operation sequences, including vibration acceleration and resonant frequency. An objective function is constructed with the goal of minimizing the total vibration energy. Constraints of no resonance and no vibration superposition are set. The no resonance constraint is that the vibration frequency is not within the preset range of the cabinet's natural frequency, and the no vibration superposition constraint is that the interval between adjacent operation vibration peaks is not less than a preset duration. The optimal operation time sequence is solved by a genetic algorithm, and the execution start time of each robot is output. The robots are controlled to execute the parcel retrieval operation in sequence to ensure the safety of the cabinet structure and the efficient and orderly retrieval.

[0060] This application's embodiments employ a physical process model for parcel delivery into the low-altitude parcel locker based on its hardware structure. Multi-dimensional real-time data is collected during parcel delivery, and the physical model is used to verify compliance during the delivery process. Based on the verification results, a state model for the corresponding locker compartment is established. When retrieving a parcel, a safe retrieval instruction is generated using this model, and the instruction and credential are sent to a robot to complete the automatic retrieval. These technical means solve the technical problem of insufficient overall safety and controllability in existing low-altitude parcel locker automatic retrieval management, achieving the technical effect of improving the safety and controllability of the entire low-altitude parcel locker automatic retrieval process.

[0061] In the above text, refer to Figure 1 This paper describes in detail an automatic parcel pickup management method based on a low-altitude parcel locker according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes an automatic parcel pickup management system based on a low-altitude parcel locker according to an embodiment of the present invention.

[0062] An automatic parcel retrieval management system based on a low-altitude parcel locker, according to an embodiment of the present invention, addresses the technical problem of insufficient overall safety and controllability in existing automatic parcel retrieval management systems for low-altitude parcel lockers, thereby improving the overall safety and controllability of the automatic parcel retrieval process. The system includes: a locker entry process mechanism model construction module 10, a dynamic sensing data acquisition module 20, a physical logic consistency verification module 30, a locker state mechanism model construction module 40, a retrieval instruction generation module 50, and an automatic locker opening and retrieval module 60.

[0063] The system comprises the following modules: Module 10 for constructing a mechanism model of the entry process, which is used to construct a mechanism model of the entry process based on prior knowledge of the structural parameters of the low-altitude express delivery locker; Module 20 for collecting dynamic sensing data during the entry process of express packages into the locker using the multimodal sensing unit built into the low-altitude express delivery locker when the package is delivered at the low-altitude delivery terminal; Module 30 for verifying the physical and logical consistency of the dynamic sensing data using the mechanism model of the entry process, which generates package retrieval management information; Module 40 for constructing a mechanism model of the locker state, which is used to construct a locker state mechanism model of the target locker based on the package retrieval management information and the mechanical characteristic parameters of the low-altitude express delivery locker; Module 50 for generating a package retrieval instruction based on the locker state mechanism model after the package retrieval task is triggered; and Module 60 for automatically opening and retrieving the package, which sends the package retrieval instruction and a one-time package retrieval voucher to the logistics robot to drive the automatic opening and retrieval operation.

[0064] The detailed description of the specific configuration of the cabinet entry process mechanism model construction module 10 is as follows: As mentioned above, based on the prior knowledge of the structural parameters of the low-altitude express cabinet, a cabinet entry process mechanism model is constructed. The cabinet entry process mechanism model construction module 10 may further include: a structural parameter prior knowledge acquisition unit for acquiring the cabinet compartment geometric dimensions, material properties, and sensor placement parameters of the low-altitude express cabinet to obtain the structural parameter prior knowledge; and a cabinet entry process mechanism model establishment unit for establishing the cabinet entry process mechanism model based on the structural parameter prior knowledge.

[0065] The specific configuration of the dynamic sensing data acquisition module 20 is described in detail below: As mentioned above, when a parcel is delivered at a low-altitude delivery terminal, dynamic sensing data during the parcel entry process is collected through the multimodal sensing unit built into the low-altitude parcel locker. The dynamic sensing data acquisition module 20 may further include: a pressure timing signal acquisition unit for acquiring pressure timing signals during the parcel entry process through a distributed pressure sensing unit; a trajectory posture image sequence acquisition unit for acquiring trajectory posture image sequences during the parcel entry process through a visual sensing unit; a cabinet structure micro-strain signal acquisition unit for acquiring cabinet structure micro-strain signals during the parcel entry process through a strain sensing unit; and a data cleaning unit for cleaning the pressure timing signal, the trajectory posture image sequence, and the cabinet structure micro-strain signal to obtain the dynamic sensing data.

[0066] The physical logic consistency verification module 30 is described in detail below: As mentioned above, the physical logic consistency verification of the dynamic sensing data is performed using the cabinet entry process mechanism model to generate package retrieval management information. The physical logic consistency verification module 30 may further include: a cabinet entry collision instantaneous measured speed calculation unit for parcel entry collision instantaneous measured speed calculated from the trajectory posture image sequence in the dynamic sensing data; a package measured mass calculation unit for parcel entry collision instantaneous measured mass calculated from the pressure time sequence signal in the dynamic sensing data; and a theoretical calculation unit for calculating the instantaneous measured speed of the package entry collision instantaneous measured speed. The measured speed and the measured mass of the package are input into the cabinet entry mechanism model to calculate the theoretical impact force distribution and theoretical strain response. The first comparison unit is used to perform spatiotemporal integration inversion calculation on the pressure time series signal to obtain the measured impact force distribution, and to perform a first comparison between the theoretical impact force distribution and the measured impact force distribution. The second comparison unit is used to perform a second comparison between the theoretical strain response and the cabinet structure micro-strain signal in the dynamic sensing data. The package retrieval management information generation unit is used to determine that the verification is passed when the deviations of the first comparison and the second comparison are both within the preset physical law tolerance range, and to generate the package retrieval management information accordingly.

[0067] The detailed description of the specific configuration of the cabinet state mechanism model construction module 40 is as follows: As mentioned above, based on the package retrieval management information and combined with the mechanical characteristic parameters of the low-altitude express cabinet, a cabinet state mechanism model of the target cabinet is constructed. The cabinet state mechanism model construction module 40 may further include: a model initial condition generation unit used to extract physical parameters of the express package, cabinet location information, and initial stress state after entering the cabinet from the package retrieval management information to generate initial conditions for the model; and a cabinet state mechanism model establishment unit used to establish the cabinet state mechanism model based on the initial conditions.

[0068] The detailed description of the specific configuration of the retrieval instruction generation module 50 is explained as follows: As mentioned above, after the retrieval task is triggered, a retrieval instruction is generated based on the cabinet state mechanism model. The retrieval instruction generation module 50 may further include: a simulation calculation unit for simulating the cabinet state mechanism model based on the retrieval task issued by the background management system, using a predetermined retrieval time window and environmental wind speed as boundary conditions; a simulation result acquisition unit for acquiring simulation results, including the optimal opening speed curve and angle curve of the cabinet door and the standard docking posture of the robotic arm that meet the mechanism safety constraints; and a retrieval instruction generation unit for encapsulating the optimal opening speed curve of the cabinet door, the angle curve, and the standard docking posture of the robotic arm to generate the retrieval instruction.

[0069] The detailed description of the automatic cabinet opening and retrieval module 60 is explained as follows: As mentioned above, the retrieval instruction and one-time retrieval voucher are sent to the logistics robot to drive the automatic cabinet opening and retrieval operation. The automatic cabinet opening and retrieval module 60 may further include: a binding unit for the back-end management system to bind the retrieval instruction and one-time retrieval voucher and send them to the assigned logistics robot; an identity authentication unit for the logistics robot to navigate to the low-altitude express cabinet and send the one-time retrieval voucher to the low-altitude express cabinet for identity authentication; an opening signal triggering unit for the logistics robot to adjust the end effector according to the standard docking posture of the robotic arm in the retrieval instruction after successful authentication, and to physically dock with the cabinet door mechanism to trigger the opening signal; an unlocking unit for the low-altitude express cabinet to drive the cabinet door locking mechanism to unlock after verifying that the current state meets the mechanism safety constraints in the retrieval instruction based on the cabinet state feedback in real time from the cabinet state mechanism model; and a package retrieval unit for the logistics robot to control the cabinet door to open and retrieve the package according to the optimal opening speed curve and angle curve in the retrieval instruction.

[0070] The system, after issuing the pickup instruction and one-time pickup voucher to the logistics robot to drive the automatic cabinet opening and pickup operation, may further include: a full-cycle actual dynamic data acquisition module for acquiring full-cycle actual dynamic data of the cabinet door closing process through the multimodal sensing unit after the logistics robot controls the cabinet door to close; a real-time comparison module for comparing the full-cycle actual dynamic data with the predicted dynamic data of the cabinet state mechanism model for the same closing process in real time; an operation compliance judgment module for determining that the pickup operation is compliant if the comparison deviation is within the preset safety tolerance range, and confirming the prediction accuracy of the cabinet state mechanism model under the current working condition; and a reverse calibration module for determining that the model prediction is inaccurate if the comparison deviation exceeds the safety tolerance range, and performing reverse calibration and optimization updates on the cabinet state mechanism model based on the full-cycle actual dynamic data.

[0071] The system may further include: a scheduling operation module for the background management system to call the cabinet state mechanism model, perform simulation prediction to obtain scheduling decision basis, and perform at least one of the following scheduling operations: marking parcels whose displacement risk value predicted by the cabinet state mechanism model is higher than a threshold as high priority, and assigning priority retrieval tasks to the logistics robots; when a new parcel needs to be stored, querying the structural fatigue value predicted by the cabinet state mechanism model for each empty cabinet, and prioritizing the allocation of the new parcel to the cabinet with the lowest current fatigue value; when multiple logistics robots need to retrieve parcels from different cabinets of the same low-altitude express cabinet, simulating the overall vibration response of the cabinet under different operation sequences of each robot based on the cabinet state mechanism model, calculating a multi-task execution time sequence to avoid overall vibration superposition or resonance, and controlling the robots to perform retrieval operations sequentially according to the multi-task execution time sequence.

[0072] The automatic parcel pickup management system based on low-altitude parcel lockers provided in this embodiment of the invention can execute the automatic parcel pickup management method based on low-altitude parcel lockers provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0073] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0074] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for automatic package pickup management based on low-altitude express lockers, characterized in that, The method includes: Based on prior knowledge of the structural parameters of low-altitude express delivery lockers, a mechanism model of the locker entry process is constructed. When delivering parcels at low-altitude delivery terminals, dynamic sensing data during the parcel entry process is collected through the multimodal sensing unit built into the low-altitude parcel locker. The physical and logical consistency of the dynamic sensing data is verified using the cabinet entry process mechanism model to generate item retrieval management information. Based on the package pickup management information and combined with the mechanical characteristic parameters of the low-altitude express cabinet, a cabinet state mechanism model of the target cabinet is constructed. After the pickup task is triggered, a pickup instruction is generated based on the cabinet state mechanism model; The pickup instruction and one-time pickup voucher are sent to the logistics robot to drive the automatic cabinet opening and pickup operation.

2. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, Based on prior knowledge of the structural parameters of low-altitude express delivery lockers, a mechanism model of the locker insertion process is constructed, including: Obtain the cabinet compartment geometry, material properties, and sensor placement parameters of the low-altitude express delivery locker to obtain prior knowledge of the structural parameters; Based on the prior knowledge of the structural parameters, a mechanism model of the cabinet entry process is established.

3. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, When delivering packages at low-altitude delivery terminals, the multimodal sensing unit built into the low-altitude express locker collects dynamic sensing data during the package's entry into the locker, including: The pressure timing signal during the process of parcels being placed in the locker is collected through a distributed pressure sensing unit. The visual sensing unit collects a sequence of trajectory and posture images of the package during the process of it being placed into the locker. The strain sensing unit collects micro-strain signals of the cabinet structure during the process of parcels being placed into the cabinet. After cleaning the pressure time-series signal, the trajectory attitude image sequence, and the cabinet structure micro-strain signal, the dynamic sensing data is obtained.

4. The automatic package pickup management method based on low-altitude express cabinets as described in claim 3, characterized in that, The physical and logical consistency of the dynamically sensed data is verified using the aforementioned cabinet entry process mechanism model to generate item retrieval management information, including: The instantaneous measured speed of the package upon collision with the cabinet is obtained by parsing and calculating the trajectory and posture image sequence in the dynamic sensing data. The measured mass of the express shipment is obtained by analyzing and calculating the pressure time-series signal in the dynamic sensing data. The measured instantaneous velocity of the collision into the cabinet and the measured mass of the express package are input into the cabinet entry process mechanism model to calculate the theoretical impact force distribution and theoretical strain response. The measured impact force distribution is obtained by performing spatiotemporal integration inversion on the pressure time series signal, and the theoretical impact force distribution is compared with the measured impact force distribution in the first comparison. A second comparison is made between the theoretical strain response and the micro-strain signal of the cabinet structure in the dynamic sensing data; When the deviations of the first comparison and the second comparison are both within the preset physical law tolerance range, the verification is deemed to have passed, and the item retrieval management information is generated accordingly.

5. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, Based on the package pickup management information and combined with the mechanical characteristic parameters of the low-altitude express locker, a locker state mechanism model of the target locker is constructed, including: From the package retrieval management information, extract the physical parameters of the package, the location information of the locker, and the initial conditions for generating the initial stress state model after the package is placed in the locker; Based on the initial conditions of the model, a state mechanism model of the cabinet is established.

6. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, After the pickup task is triggered, a pickup instruction is generated based on the cabinet state mechanism model, including: Based on the pickup task issued by the back-end management system, the scheduled pickup time window and environmental wind speed are used as boundary conditions to input the cabinet state mechanism model for simulation calculation. Obtain simulation results, including the optimal opening speed curve and angle curve of the cabinet door and the standard docking posture of the robotic arm that meet the mechanistic safety constraints; The optimal opening speed curve of the cabinet door, the angle curve, and the standard docking posture of the robotic arm are encapsulated to generate the part retrieval command.

7. The automatic package pickup management method based on low-altitude express cabinets as described in claim 6, characterized in that, The pickup instruction and one-time pickup voucher are sent to the logistics robot to drive the automatic cabinet opening and pickup operation, including: The back-end management system binds the pickup instruction and the one-time pickup voucher and sends them to the assigned logistics robot; The logistics robot navigates to the low-altitude express cabinet and sends the one-time pickup certificate to the low-altitude express cabinet for identity authentication. After authentication, the logistics robot adjusts its end effector according to the standard docking posture of the robotic arm in the pickup instruction, and physically docks with the cabinet door mechanism to trigger an opening signal. The low-altitude express cabinet is based on the cabinet status feedback in real time from the cabinet status mechanism model. After verifying that the current status meets the mechanism safety constraint conditions in the package retrieval instruction, the cabinet door locking mechanism is driven to unlock. The logistics robot controls the cabinet door to open and retrieve the package based on the optimal opening speed curve and angle curve in the pickup instruction.

8. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, After issuing the pickup instruction and one-time pickup voucher to the logistics robot to drive the automatic cabinet opening and pickup operation, the system also includes: After the logistics robot closes the cabinet door, the multimodal sensing unit collects the actual dynamic data of the entire cycle of the cabinet door closing process. The actual dynamic data of the entire cycle is compared in real time with the predicted dynamic data of the cabinet state mechanism model for the same closed process; If the comparison deviation is within the preset safety tolerance range, the current item retrieval operation is deemed compliant, and the prediction accuracy of the cabinet state mechanism model under the current working conditions is confirmed. If the comparison deviation exceeds the safety tolerance range, it is determined that the model prediction is inaccurate. Based on the actual dynamic data of the whole cycle, the cabinet state mechanism model is reverse calibrated and optimized.

9. The automatic package pickup management method based on low-altitude express cabinets as described in claim 1, characterized in that, Also includes: The backend management system calls the cabinet state mechanism model, performs simulation prediction to obtain scheduling decision basis, and performs at least one of the following scheduling operations: Packages whose displacement risk value predicted by the cabinet state mechanism model is higher than the threshold are marked as high priority and assigned priority pickup tasks to the logistics robot. When a new package needs to be stored, the structural fatigue level predicted by the cabinet state mechanism model corresponding to each available cabinet is queried, and the new package is preferentially assigned to the cabinet with the lowest current fatigue level. When multiple logistics robots need to retrieve packages from different compartments of the same low-altitude express cabinet, the overall vibration response of the cabinet is simulated under different operation sequences of each robot based on the cabinet state mechanism model. The multi-task execution time sequence is calculated to avoid the superposition or resonance of overall vibration, and the robots are controlled to perform the package retrieval operation in sequence according to the multi-task execution time sequence.

10. An automatic parcel pickup management system based on low-altitude express lockers, characterized in that, The system is used to implement the automatic package pickup management method based on low-altitude express cabinets as described in any one of claims 1-9, the system comprising: The module for constructing the mechanism model of the cabinet entry process is used to construct the mechanism model of the cabinet entry process based on the prior knowledge of the structural parameters of the low-altitude express cabinet. The dynamic sensing data acquisition module is used to collect dynamic sensing data of the express delivery package during the delivery process at the low-altitude delivery terminal by using the multimodal sensing unit built into the low-altitude express cabinet. The physical-logical consistency verification module is used to perform physical-logical consistency verification on the dynamic sensing data using the cabinet entry process mechanism model, and generate item retrieval management information. The cabinet state mechanism model construction module is used to construct a cabinet state mechanism model of the target cabinet based on the package retrieval management information and the mechanical characteristic parameters of the low-altitude express cabinet. The item retrieval instruction generation module is used to generate an item retrieval instruction based on the cabinet state mechanism model after the item retrieval task is triggered. The automatic cabinet opening and retrieval module is used to send the retrieval instruction and one-time retrieval voucher to the logistics robot, driving it to perform the automatic cabinet opening and retrieval operation.

Images