Warehouse unmanned aerial vehicle-mounted RFID scanning automated warehouse method and system

Abstract

The application discloses an automatic warehouse inventory method and system based on a warehouse-mounted RFID scanning unmanned aerial vehicle, and belongs to the technical field of warehouse logistics automation. The method comprises the following steps: a warehouse control server plans an unmanned aerial vehicle inventory route containing a scanning trigger point based on a warehouse three-dimensional model; the unmanned aerial vehicle autonomously flies along the route, and an RFID scanning module is automatically started by a trigger control unit to read a label when flying over a storage location trigger area; read data is real-timely returned to the server for binding and processing. The application realizes the full-process automation of the stereoscopic warehouse inventory operation, replaces the inefficient and dangerous manual inventory, utilizes the maneuverability of the unmanned aerial vehicle in combination with the non-contact reading of the RFID, greatly improves the inventory efficiency, safety and accuracy, has the advantages of flexible deployment, strong adaptability, support for multi-machine cooperation and the like, and is suitable for the inventory management of various large stereoscopic warehouses.

Description

Technical Field

[0001] This invention relates to the field of automated warehousing and logistics technology, and in particular to an automated inventory counting method and system using drones equipped with RFID scanning in warehouses. Specifically, it relates to a method and system for fully automated, efficient, and secure inventory counting in large automated warehouses based on drones equipped with RFID scanning modules. More specifically, this invention belongs to the intersection of intelligent warehousing, drone automatic control, and Internet of Things (IoT) data acquisition and processing technologies. It focuses on solving the problem of how to integrate RFID reading and writing devices into autonomous drone platforms in high-density automated warehousing environments to achieve accurate, continuous, and contactless automated inventory data collection, and deeply integrates the collection process with a warehouse management system (WMS) to form a closed-loop inventory information update and management process. Background Technology

[0002] With the rapid development of e-commerce, intelligent manufacturing, and modern logistics, large-scale automated warehouses have become the mainstream form of warehousing systems. These warehouses typically feature high rack heights (ranging from 5 to 20 meters), dense storage locations, a wide variety of goods, and high inbound and outbound frequencies. Timely and accurate inventory information is crucial for enterprises to optimize inventory structure, improve warehousing efficiency, and reduce operating costs.

[0003] Currently, inventory counting in large automated warehouses mainly relies on manual operation. The common method involves operators driving lifting vehicles or elevator platforms, moving row by row and column by column to each storage location, and using handheld barcode scanners or RFID readers to scan the goods. This traditional inventory counting method has the following significant drawbacks:

[0004] Inefficiency: Manually operating equipment, locating storage locations, and scanning each item are extremely time-consuming. For a medium-sized automated warehouse, a complete inventory check often takes several days or even weeks, severely impacting normal warehouse operations and failing to meet the real-time or near-real-time requirements for inventory data in high-turnover scenarios.

[0005] Significant safety hazards exist: Operators need to work on the lifting platform at heights (several meters to over ten meters) for extended periods, posing major safety risks such as equipment falling, collisions with shelves, and personnel falling. Additionally, when manually scanning the shelves, there is a possibility of being injured by falling goods.

[0006] Inventory accuracy is difficult to guarantee: manual operation is prone to omissions, duplicates, or incorrect scans due to fatigue, obstructed vision, or operational errors. Data recording may also introduce errors during subsequent manual entry, reducing the reliability of the inventory results and causing discrepancies with actual inventory.

[0007] High labor costs and reliance on experience: Training professional inventory personnel is costly and the workload is heavy. Against the backdrop of rising labor costs and a shortage of human resources, the sustainability of this model faces challenges.

[0008] Low level of automation and informatization: The existing inventory process has limited integration with the warehouse management system (WMS) or warehouse control system (WCS), making it difficult to achieve automatic issuance of inventory tasks, full monitoring of the execution process, and real-time seamless transmission and comparison of inventory data.

[0009] In recent years, while some research has attempted to introduce automation into the inventory process—such as deploying fixed RFID reader networks on shelves or using automated guided vehicles (AGVs) equipped with scanning devices on tracks—fixed reader networks are extremely expensive to deploy, difficult to cover all storage locations, and lack flexibility. Track-based AGVs are limited by fixed paths, cannot flexibly adapt to the scanning needs of shelves at different heights, and require significant engineering work and high costs to modify existing warehouse infrastructure. Furthermore, existing semi-automated solutions suffer from data synchronization delays, insufficient system robustness, and poor adaptability to complex dynamic environments. For example, fixed reader networks may experience blind spots due to goods obstructing the view or metal interference; track-based AGVs require manual intervention to replan when faced with temporary storage or path changes, failing to achieve truly fully automated and intelligent responses.

[0010] Therefore, the industry urgently needs an automated inventory management solution for automated warehouses that can balance high efficiency, high security, high accuracy, high automation, and easy deployment. Drone technology, due to its flexible maneuverability and controllable flight altitude and path, is showing increasing potential in indoor warehousing environments. However, how to effectively combine drones with RFID scanning technology to solve the problem of stable, reliable, and continuous "in-flight scanning" in complex and dense automated warehouse environments, including a series of technical challenges such as equipment integration, precise positioning triggering, real-time data transmission, and flight safety obstacle avoidance, has yet to be addressed with a mature and comprehensive systematic solution. In particular, existing technologies lack a complete technical solution that can perform refined flight path planning based on a precise 3D model of the warehouse and dynamically match high-precision real-time positioning with preset trigger areas to achieve automatic and precise start and stop of RFID scanning by drones during high-speed flight. Furthermore, existing publicly available technical solutions do not provide systematic solutions for key aspects such as adaptive retry mechanisms for failed reads, airspace management and conflict resolution during multi-drone collaborative operations, and real-time deduplication verification and binding processing of massive amounts of RFID data. Summary of the Invention

[0011] To address the shortcomings of existing technologies, the purpose of this invention is to provide an automated inventory method and system for warehouses using drones equipped with RFID scanning. The core of this method is to utilize an indoor drone with autonomous flight and obstacle avoidance capabilities, equipped with a customized small RFID scanning module. Through a pre-planned or real-time generated refined flight path, the drone flies along the warehouse aisles. When it flies over a specific area of ​​each storage location, the trigger control unit automatically initiates RFID scanning, achieving continuous, non-contact inventory counting by "scanning as it flies over" and synchronizing the collected data to the central control system in real time.

[0012] This invention employs a collaborative architecture of "cloud planning - edge execution - real-time feedback," placing complex global path planning and inventory data processing tasks on the warehouse control server (cloud), while real-time flight control, obstacle avoidance decision-making, and scan triggering tasks are handled by the UAV's onboard system (edge). This achieves rational allocation of computing resources and maximizes system efficiency. Crucially, this invention creatively proposes the concepts of "virtual scan points" and "3D scan trigger areas," transforming discrete inventory location tasks into responses to a series of spatial triggering events during UAV flight along a continuous path. This enables streamlined and highly automated inventory operations.

[0013] The above-mentioned objective of this invention is achieved through the following technical solutions:

[0014] This invention provides an automated inventory counting method using RFID scanning via drones in a warehouse, comprising the following steps:

[0015] Step S1: The warehouse control server generates the drone's inventory flight path based on the three-dimensional spatial information of the target inventory area. The flight path includes multiple scanning trigger points associated with the inventory location and their corresponding scanning trigger area information.

[0016] Step S2: The UAV loads the inventory flight route, takes off from the starting point and flies autonomously along the route, while using onboard obstacle avoidance sensors for real-time environmental perception and dynamic obstacle avoidance.

[0017] Step S3: During flight, the trigger control unit continuously acquires the real-time location information of the UAV and compares it with the scanning trigger area information; when it is determined that the UAV has entered the scanning trigger area of ​​any cargo location, the RFID scanning module on the UAV is automatically triggered to read the RFID signal of the cargo location.

[0018] Step S4: The RFID scanning module will send the RFID tag data it reads to the warehouse control server through the wireless data transmission unit;

[0019] Step S5: The warehouse control server receives the RFID tag data, binds and processes it with the corresponding storage location information, and updates the inventory data;

[0020] Step S6: After completing the flight path, the UAV automatically returns and lands.

[0021] Furthermore, the method is a complete closed-loop control system. After updating the inventory data in step S5, the warehouse control server can generate an inventory discrepancy report and trigger subsequent operation instructions based on this report, such as generating a stock transfer task order or triggering a review process. This allows the inventory results to be directly applied to the optimization of inventory management, forming a closed loop from data collection to decision execution.

[0022] According to one embodiment of the present invention, in step S1, generating the inventory flight path specifically includes: generating a horizontal continuous flight path for each shelf height that needs to be inventoried based on the position, height and coordinates of the storage location in the three-dimensional digital model of the warehouse, combined with the flight performance parameters and safe obstacle avoidance distance of the UAV, and setting a virtual scanning point at a preset distance directly in front of each storage location to be inventoried on the path, wherein the virtual scanning point is associated with a scanning trigger area in three-dimensional space.

[0023] The preset distance is determined based on the effective reading distance of the RFID scanning module, the positioning accuracy of the UAV, and its flight stability, and is typically between 0.5 meters and 2 meters. Flight path planning also needs to consider the UAV's turning radius and maximum climb rate to ensure a smooth, flyable path with optimal energy consumption. For ultra-large warehouses, the planning module supports dividing the overall flight path into multiple sub-task segments to accommodate the UAV's endurance limitations.

[0024] According to one embodiment of the present invention, the scanning triggering area is a spherical area, a cubic area, or a fan-shaped area defined according to the direction of the shelf opening, with the virtual scanning point as the center and a preset length as the radius.

[0025] Preferably, to adapt to common shelving structures and improve triggering accuracy, the scanning trigger area is often defined as a fan-shaped area facing the shelving opening, with its central axis perpendicular to the shelving surface and an angle between 60° and 120°. This shape better matches the flight trajectory of drones approaching the storage location from the aisle, reducing false or missed triggers. The geometric parameters of the trigger area (such as radius and angle) can be differentiated according to the label density and the material of the goods (whether they obstruct the radio frequency signal) in different storage locations.

[0026] According to one embodiment of the present invention, in step S3, the step of automatically triggering the RFID scanning module to read further includes: after triggering the reading, receiving reading feedback from the RFID scanning module; if the reading feedback is a failure state, then according to a preset retry strategy, controlling the UAV to perform at least one of the following operations: hovering, adjusting pose, or adjusting RFID reading parameters, and then triggering the reading again.

[0027] The retry strategy is a pre-configured or dynamically issued set of strategies by the server, which may include, but is not limited to: immediate retry after hovering in place, retry after slight lateral or forward / backward movement, retry after increasing the RFID reader's transmission power, retry after switching the reading frequency, and performing a small-range hovering scan. The trigger control unit records the result of each retry. If the maximum number of retries is reached and the retry still fails, the location is marked as "requiring manual verification" and the environmental information at the time of failure (such as position and attitude) is recorded. Then, the drone is controlled to continue executing subsequent routes to ensure the continuity of the overall operation.

[0028] The directional antenna is fixed to the UAV body via an adjustable or fixed mounting bracket, ensuring that its main beam direction remains substantially aligned with the shelf array to be scanned during flight. The mounting bracket can be designed with vibration damping capabilities to suppress the impact of UAV rotor vibration on antenna performance. The RFID reader connects to the UAV flight control system via a standard communication interface (such as UART, CAN, USB, or Ethernet), following agreed-upon protocols for command and data transmission. If the trigger control unit is implemented as an independent module, its core can be an embedded microprocessor running dedicated trigger judgment and communication management firmware.

[0029] According to one embodiment of the present invention, the RFID scanning module includes a miniaturized UHF RFID reader and a directional antenna, wherein the main radiation direction of the directional antenna is toward the shelf on one side of the flight path; the trigger control unit is integrated into the flight controller of the UAV or is connected to the flight controller as an independent module via a communication interface.

[0030] This invention also provides an automated inventory system for warehouses using drones equipped with RFID scanning, for implementing the automated inventory method for warehouses using drones equipped with RFID scanning described in the above embodiments. The system includes:

[0031] The warehouse control server is equipped with a 3D warehouse model, a cargo location coordinate database, and a flight path planning module. It is used to generate inventory flight paths and receive and process inventory data.

[0032] At least one drone, including the drone body, flight control system, positioning module, obstacle avoidance sensor and power supply unit, is used to fly autonomously according to the inventory flight route;

[0033] An RFID scanning module, fixed to the drone body, includes an RFID reader and an antenna, and is used to read RFID tag information on the cargo location after being triggered;

[0034] The trigger control unit is communicatively connected to the flight control system of the UAV and the RFID scanning module, and is used to automatically control the start and stop of the RFID scanning module according to the matching result of the real-time position of the UAV and the preset scanning trigger area.

[0035] A data transmission unit, installed on the UAV, is used to establish a communication link between the UAV and the warehouse control server to transmit flight route instructions and inventory data.

[0036] The system is a layered distributed system. The warehouse control server constitutes the command and control layer, responsible for macro-level task management and data analysis; the UAV and its onboard systems (flight control, RFID module, trigger control unit, etc.) constitute the execution layer, responsible for specific spatial movement and data acquisition operations; the wireless communication network constitutes the connection layer, ensuring reliable interaction between commands and data. The layers communicate with each other through well-defined API interfaces and data protocols.

[0037] According to one embodiment of the present invention, the route planning module of the warehouse control server is specifically used to: generate a route file containing a series of ordered waypoints based on the warehouse 3D model, the performance constraints of the UAV and the inventory task range, wherein an associated virtual scanning point and a 3D trigger area centered on each inventory location are generated; the route file is sent to the UAV through the data transmission unit.

[0038] The route planning module employs rule-based algorithms or path search algorithms (such as the application of Dijkstra's algorithm in 3D space) for path calculation. It considers not only the shortest path but also factors such as flight safety (e.g., avoiding pillars and ventilation ducts), scanning efficiency (e.g., reducing unnecessary climbs and descents), and energy balance. The generated route file uses a structured data format (e.g., JSON, XML, or a custom binary format), containing a waypoint sequence, coordinates of each waypoint, expected speed, action commands (e.g., "trigger scan at this point," "adjust altitude at this point"), and associated cargo location IDs and trigger area descriptions.

[0039] According to one embodiment of the present invention, the trigger control unit is configured to: acquire real-time pose data of the UAV provided by the flight control system at a fixed frequency, calculate the spatial distance between the UAV and each virtual scanning point in the flight path file, and generate a trigger command and send it to the RFID scanning module when the distance is less than a preset trigger threshold.

[0040] The frequency of acquiring real-time pose data is typically between 10Hz and 100Hz to ensure timely triggering. Euclidean distance formulas can be used to calculate spatial distance. To improve computational efficiency, the trigger control unit maintains a "trigger list," performing distance calculations only on virtual scan points within a certain range of the drone's current location. The preset trigger threshold is usually slightly smaller than the geometric boundary value of the trigger area to provide a buffer and prevent repeated triggering due to minor positioning fluctuations. The trigger command may include information such as the target cargo location ID and desired RFID reading parameters (e.g., power level).

[0041] According to one embodiment of the present invention, the system further includes at least one UAV nest, the nest being provided with a charging interface and a data interface for automatically charging the landed UAV and for retransmitting cached data via a wired connection; the warehouse control server further includes a multi-UAV scheduling module for performing task partitioning, route allocation, and conflict coordination when there are multiple UAVs.

[0042] The drone nest is a smart infrastructure that, in addition to charging and data interfaces, integrates functions such as drone status self-checking, simple cleaning, and hot-swappable battery replacement. The multi-drone scheduling module is responsible for breaking down large-scale inventory tasks into multiple sub-tasks, assigning each drone a non-overlapping or conflicting flight area and time window. Conflict coordination strategies include spatial isolation (assigning different lanes), temporal isolation (staggered entry into public areas), priority settings, and dynamic avoidance based on real-time communication. By monitoring the status of all drones (location, battery level, task progress), the scheduling module can dynamically reallocate tasks; for example, when a drone malfunctions and exits, its unfinished tasks are automatically assigned to other idle or soon-to-be-idle drones.

[0043] According to one embodiment of the present invention, the warehouse control server further includes a data receiving and processing module for parsing, verifying, deduplicating, and binding the received RFID tag data with the storage location information, and further includes a report generation module for comparing the processed actual inventory data with the system's book inventory after the inventory task is completed, identifying differences, and generating an inventory report.

[0044] The data receiving and processing module implements a series of data cleaning and association rules: parsing the raw RF data packets to obtain EPC codes; verifying whether the EPC codes conform to specifications; for multiple read results reported by the same storage location within a short period of time, deduplication is performed based on time windows and signal strength (RSSI), retaining the record with the strongest signal or the first valid read; finally, the tag data is bound to the specific physical storage location through the mapping relationship between task ID and storage location ID. The report generation module can generate reports in various formats and dimensions, including summary reports, detailed discrepancy reports, inventory efficiency statistics reports (such as inventory time, scan success rate), and visualization reports (such as highlighting discrepancy storage locations on the warehouse floor plan).

[0045] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects:

[0046] Revolutionary improvement in inventory efficiency: Drones can fly at speeds of 1-3 meters per second and achieve "in-flight scanning," eliminating the need to stop at each storage location, making inventory operations much faster than manual methods. Practice shows that overall inventory efficiency can be 5-10 times higher than traditional manual inventory, reducing the time for full-warehouse inventory checks in large warehouses from days to hours, or even less, greatly supporting high-frequency, periodic inventory needs.

[0047] The safety of operations has been fundamentally improved: the risks of personnel working at heights and having close contact with heavy shelves have been completely eliminated, fundamentally eliminating the possibility of related safety accidents and meeting the highest standards of modern warehouse safety production.

[0048] High accuracy and reliability of inventory checks: Automated scanning avoids subjective errors and oversights caused by human fatigue. The contactless, batch reading characteristics of RFID technology, combined with a precise triggering mechanism, ensure that every storage location is effectively covered. Data is automatically uploaded to the system, avoiding manual entry errors and guaranteeing consistency between records and actual inventory.

[0049] High degree of automation and intelligence: It achieves full automation from mission planning, autonomous flight, automatic scanning to data transmission and processing, significantly reducing reliance on manpower and labor intensity. The system can be deeply integrated with the upper-level WMS / WCS to realize closed-loop management and dynamic updating of inventory information.

[0050] The system boasts high flexibility and scalability: Strong environmental adaptability: By replanning flight routes through software, it can easily adapt to automated warehouses of different layouts and sizes without requiring modifications to the warehouse's physical structure. Modular design: The RFID scanning module adopts standardized interfaces and a lightweight design, facilitating rapid integration and replacement on different models of compliant drones. Support for multi-drone collaboration: The system architecture supports unified task scheduling and monitoring of multiple drones, further shortening inventory time in large-scale warehouses through parallel operations.

[0051] Significant cost-effectiveness: While there is some initial hardware investment, in the long run, it saves substantial amounts of labor, training, and safety risk management costs. The efficiency improvements leading to optimized warehouse operations (such as reduced inventory backlog and faster turnover) offer even more substantial benefits. Compared to deploying fixed RFID networks or track systems, this solution has lower deployment and maintenance costs and greater flexibility.

[0052] The real-time performance of data is greatly enhanced: By transmitting data back in real time via wireless network, managers can monitor the progress of inventory counts and discover inventory discrepancies in near real time, supporting rapid decision-making and response, and changing the lagging model of traditional post-inventory reconciliation.

[0053] The system is robust and fault-tolerant: Through mechanisms such as local route storage, offline caching, adaptive retries, and multi-machine backup, the system can cope with abnormal situations such as temporary network interruptions, single device failures, and difficulties in reading individual tags, ensuring the smooth completion of the overall inventory task.

[0054] It provides a wealth of quantifiable management metrics: the system automatically records detailed process data for each inventory count task (flight trajectory, scanning time point, read success rate, etc.), providing data support for continuous optimization of warehouse operations (such as adjustment of storage location layout and optimization of inventory count frequency).

[0055] It promotes the digitalization and twinning of warehousing: This invention relies on a precise 3D model of the warehouse, and its operation continuously generates data that is precisely associated with the physical storage locations, which strongly promotes and supports the construction and application of warehouse digital twin systems. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the overall structure of the UAV equipped with the RFID scanning module of the present invention.

[0057] Figure 2 This is a schematic diagram of the flight path and cargo location scanning of the UAV in the automated warehouse of the present invention.

[0058] Figure 3 This is a system workflow diagram of the present invention.

[0059] Reference numerals: 1. UAV; 11. UAV body; 12. RFID reader; 13. Antenna. Detailed Implementation

[0060] To further clarify the technical solutions and advantages of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention. 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. Those skilled in the art will understand that various changes, modifications, substitutions, and variations can be made to the embodiments without departing from the spirit and scope of the present invention.

[0061] In the description of this application, it should be noted that the terms "upper," "lower," "inner," "outer," "top / bottom," etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0062] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installed," "equipped with," "sleeved / connected," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0063] It should be noted that the "warehouse control server" involved in this invention can be a single physical server, a cluster of multiple servers, or a cloud server instance. "Unmanned aerial vehicle" refers to an aircraft capable of stable hovering and autonomous flight in an indoor environment, typically a multi-rotor drone, but not limited to this. "RFID scanning module" specifically refers to a reader / writer device and its antenna combination operating in the UHF band and complying with relevant radio management regulations. In the context of this invention, "real-time" can refer to milliseconds, seconds, or minutes, depending on the processing stage; its core meaning lies in the significant reduction of information acquisition and processing latency compared to traditional manual or semi-automatic methods.

[0064] Reference Figures 1-3 This invention discloses an automated inventory system for warehouses using drones equipped with RFID scanning, comprising:

[0065] The UAV platform 1 utilizes a multi-rotor (e.g., quadcopter, hexcopter) indoor UAV 1, integrating a high-precision flight control system, a positioning module (e.g., UWB, visual odometry, laser SLAM), and obstacle avoidance sensors in multiple directions (e.g., LiDAR, depth camera, ultrasonic sensors) such as forward and downward. The flight control system includes a flight control motherboard, inertial measurement unit (IMU), and barometer, enabling precise control of the UAV's attitude, altitude, and position. The positioning module employs multi-source fusion positioning technology, such as using UWB (Ultra-Wideband) to provide absolute position anchors, combined with high-frequency relative displacement data from visual odometry or laser SLAM (Simultaneous Localization and Mapping), and fused using filtering algorithms (e.g., Kalman filtering) to output real-time pose information with centimeter-level accuracy. The obstacle avoidance sensor array covers multiple directions (forward, backward, left, right, up, and down) of the UAV, constructing a dynamic safety perception sphere to ensure effective avoidance of static and dynamic obstacles in both autonomous flight and manual intervention modes.

[0066] RFID scanning module: A miniaturized, lightweight, and low-power module specifically designed for use with UAV1. It mainly includes:

[0067] RFID Reader 12: Operating at a frequency compliant with warehouse RFID tag standards (e.g., UHF), its size, weight, and power consumption are optimized, typically no larger than 15cm × 10cm × 5cm, weighing no more than 500g, and consuming no more than 10W. This reader supports mainstream air interface protocols such as EPCglobal Class 1 Gen 2 and features multi-tag anti-collision reading capabilities. Its transmission power is software-adjustable to adapt to reading requirements at different distances and in various environments. The reader receives commands from the trigger control unit via a communication interface (e.g., USB or serial port) and encapsulates the data into a specified data packet for output after reading.

[0068] Directional antenna 13: Mounted on UAV 1, its main lobe of the radiation pattern faces the side of the shelf to be scanned to enhance the reception of target location tag signals, suppress interference from other directions, and improve the success rate and range of reading. The antenna type can be a panel antenna, dipole antenna, etc. Its installation position and angle are carefully designed, typically located on the side or lower front of the UAV fuselage. A mechanical structure ensures that, under normal UAV flight attitude, the antenna's main beam direction is substantially perpendicular to the flight path and points towards the shelf row. In some embodiments, the antenna can be dynamically adjusted at small angles via a servo mechanism to fine-tune the beam direction during flight, further optimizing reading performance.

[0069] Trigger Control Unit: This unit acts as a bridge between the UAV 1 flight control system and the RFID scanning module. It receives real-time position and attitude information from the flight control system and compares it with a preset cargo location coordinate database or real-time environmental perception data. When it determines that UAV 1 has entered a valid scanning trigger area of ​​a cargo location (e.g., a spherical space with a radius of 0.5-1.5 meters centered on the cargo location, or a sector defined according to the rack structure), this unit automatically sends a start scanning command to the RFID reader 12. After scanning is complete, a stop command is sent to save power. The core logic of the trigger control unit is to continuously execute a "perception-judgment-execution" loop. It subscribes to high-frequency pose topics from the flight control system and obtains the geometric descriptions of all virtual scanning points and their trigger areas from the loaded flight path file. In each calculation cycle, it traverses all untriggered virtual scanning points and calculates the distance or spatial relationship between the current UAV position and each point. Once the triggering conditions are met (such as entering a sector area), it immediately generates a trigger event containing the target point ID and possible read parameters, sends it to the RFID reader via the communication bus, and marks the point as "triggered" to prevent repeated actions. Simultaneously, it listens for feedback from the RFID reader and decides whether to initiate a retry process based on the feedback.

[0070] Data transmission unit: Responsible for transmitting tag data (EPC code, cargo information, etc.) read by the RFID scanning module. It typically uses an onboard wireless communication module (such as Wi-Fi 6, 5G CPE, or industrial wireless private network) to upload data in real time to an access point deployed within the warehouse, and then transmits it to the warehouse control server (or WCS / WMS server). To ensure reliability, UAV 1 can also have local data caching capabilities, temporarily storing data during network interruptions and retransmitting it after the connection is restored. The data transmission unit supports dual-link backup, for example, simultaneously supporting Wi-Fi and 4G / 5G cellular networks, automatically switching when the main link signal is weak or interrupted. Data packet transmission uses reliable protocols (such as TCP or UDP-based application-layer reliable protocols) and includes sequence numbers and verification information to ensure data integrity and order. Local caching typically uses onboard flash memory or SD cards, storing unsuccessfully uploaded data packets indexed by task ID and time.

[0071] Warehouse control server: This is the "brain" of the system. Its functions include:

[0072] Warehouse digital modeling: storing or generating a 3D digital model of the warehouse, including the geometric dimensions and location coordinates of all shelves, and the precise 3D coordinates of each storage location (warehouse location). The model can be derived from Building Information Modeling (BIM), a point cloud model generated by laser scanning, or a simplified model digitized from warehouse design drawings. This model serves as the "digital chassis" for the system's operation and needs to be periodically calibrated against the actual physical warehouse to ensure consistency.

[0073] Task and Flight Route Planning: Upon receiving inventory task requests, the system automatically generates the optimal inventory flight route based on the warehouse model, UAV performance parameters (turning radius, climb rate, obstacle avoidance distance, etc.), and current cargo location occupancy status. The route is typically a serpentine path, covering all aisles and cargo location height levels requiring inventory. The planning engine is a core software module. It first filters all cargo locations to be inventoried based on the task scope. Then, based on graph theory or heuristic algorithms, it finds one or a set of paths within the feasible airspace of the warehouse's three-dimensional space that can efficiently traverse all associated virtual scan points. Planning must strictly adhere to airspace constraints (such as no-fly zones and low-lying areas), dynamic constraints, and safety constraints. For multi-UAV collaborative tasks, the planning module also needs to combine with the scheduling module for collaborative path planning to avoid flight route conflicts.

[0074] Data reception and processing: Receive RFID scan data from UAV 1, bind it to the corresponding cargo location information, deduplicate, verify, and update the inventory database. The data processing pipeline includes: data access (listening to network ports or message queues), format parsing, business logic verification (e.g., checking the validity of the task ID and whether the UAV is authorized), data cleaning (filtering invalid EPCs and deduplicating based on spatiotemporal windows), data association (binding tag EPCs to cargo location IDs in trigger events), and result persistence (writing to the inventory result database). The entire process can be stream-processed, enabling real-time data entry.

[0075] Monitoring and Scheduling: Real-time monitoring of Drone 1's flight status, location, battery level, and scanning progress, supporting multi-drone collaborative scheduling. The monitoring interface graphically displays the warehouse map, drone's real-time location (icon), flight trajectory, and inventory status (e.g., inventoried, uninvented, discrepancies). The scheduling module dynamically allocates and adjusts drone tasks based on task requirements, drone status, and real-time warehouse conditions (e.g., temporary occupancy of certain aisles). It is responsible for issuing commands for drone takeoff, landing, mission pause, and emergency recall.

[0076] Drone 1 Nest (Optional): Used for the automatic takeoff, landing, charging, high-speed wired data upload (during landing), and simple maintenance of Drone 1. The nest can be deployed in a fixed location within the warehouse to achieve fully automated cycle operation of Drone 1. The nest contains precise positioning and guidance devices (such as visual targets and infrared beacons) to guide the drone to a precise landing; a mechanical locking mechanism to secure the drone; charging contacts or a wireless charging module to charge the battery; a high-speed wired data interface (such as Gigabit Ethernet) for quickly downloading flight logs and cached data; environmental monitoring sensors (temperature and humidity); and a simple cleaning device (air blower). The nest itself also communicates with the warehouse control server via a network, reporting its status and the status of the drones it houses.

[0077] The typical workflow of the method is as follows:

[0078] Task Initiation and Planning: The operator initiates an inventory task on the warehouse control server interface, selecting the area to be inventoried (e.g., the entire warehouse, a specific aisle, or certain shelving rows). Based on the latest warehouse 3D model and task parameters, the server automatically generates a detailed flight path file, including a series of ordered waypoints and the expected actions for each waypoint (e.g., adjusting altitude, triggering a scan). The flight path file is distributed to the designated UAV1 via a wireless network. Task parameters may also include priority, planned execution time, whether dynamic obstacle avoidance is allowed, and retry strategy configuration. The planning process may be completed in seconds to minutes, depending on the warehouse size and computing resources.

[0079] UAV 1 Takeoff and Cruise: UAV 1 automatically takes off from its nest or designated takeoff point and flies towards the mission starting point. During flight, its flight control system controls the flight according to the flight path instructions, while obstacle avoidance sensors continuously operate, detecting and avoiding sudden obstacles (such as temporarily stacked cargo or other mobile equipment) in real time, and dynamically adjusting the local path. The flight control system follows the principle of "global path tracking as the primary method, and local obstacle avoidance correction as a secondary method." That is, it prioritizes tracking the issued flight path, and when the obstacle avoidance sensor detects an obstacle within a safe distance, it triggers the local planner to generate a temporary local path to bypass the obstacle, and then converges back to the original global flight path as quickly as possible after bypassing the obstacle. The local planning algorithm can be the artificial potential field method, the dynamic window method (DWA), etc.

[0080] Automatic Triggering and Scanning: Drone 1 flies along the planned route. The trigger control unit continuously calculates the current position of Drone 1. When it detects that Drone 1 has entered a preset "scan trigger zone" of a cargo location (e.g., by comparing the coordinates of Drone 1 with the coordinates of the cargo location, and the distance is less than a set threshold), it immediately sends a command to the RFID reader 12. The RFID reader 12 powers on, transmits a query signal through the directional antenna 13, and receives the reply signal from the RFID tags affixed to the cargo location. After decoding, it obtains the cargo information stored in the tag. A single scan process is typically completed within a few hundred milliseconds. Trigger judgment is a millisecond-level response. After receiving the command, the RFID reader performs one or more inventory operations, attempting to read all tags within its radio frequency field. The reading results (a list of successful tags or failure information) are immediately packaged.

[0081] Data transmission: The read tag data is sent back to the warehouse control server in real-time or near real-time via the data transmission unit. The server-side program associates the received data with the location ID that triggered the scan and writes it to the temporary inventory database. Transmission latency is typically between milliseconds and seconds, depending on network conditions. The data packets use a lightweight structure and contain key fields such as timestamps, drone IDs, location IDs, and a list of tag data.

[0082] Continuous operation: After completing the scan of one cargo location, UAV 1 does not need to hover and wait, but continues to fly to the next waypoint / cargo location, repeating steps 3 and 4 to form a continuous scanning workflow. This "assembly line" operation mode is the key to improving efficiency. After the UAV flies away from the previous trigger area, the trigger control unit immediately begins to prepare for the next trigger area, while the flight control smoothly guides the UAV to the next waypoint.

[0083] Mission Completion and Return: After completing the flight and scanning of all planned waypoints, UAV 1 will automatically fly back to its nest or designated landing point. During or after landing, it will fully upload flight logs and any cached data. The return-to-home decision can also be triggered based on a battery threshold. The flight control system will intelligently decide whether to continue the remaining tasks or return immediately based on the remaining battery power and return distance, ensuring the UAV's safe return.

[0084] Data Comparison and Report Generation: After receiving the inventory count data, the warehouse control server runs a data comparison algorithm to compare the actual scanned inventory information with the system's recorded inventory. It automatically identifies discrepancies (overages, shortages, misplacements, etc.) and generates a structured inventory report for management review and processing. The comparison algorithm considers various business scenarios, such as allowing multiple items to be stored in one location or having multiple labels for the same item. Once the report is generated, it can automatically notify relevant personnel via email, push notifications, or system notifications, and may trigger inventory adjustment processes within the WMS.

[0085] The technical solutions, implementation details, and application variations of the present invention in different scenarios are further illustrated below through several specific embodiments.

[0086] Example 1: Application of large-scale e-commerce warehousing centers

[0087] Reference Figures 1-3 This embodiment takes a large e-commerce regional distribution center as an example. Its automated warehouse is 150 meters long, 80 meters wide, and the shelving height is 12 meters. It has 6 floors and a total of approximately 50,000 storage locations. The aisle width is 3.5 meters.

[0088] Step S101: System Initialization and Modeling

[0089] The warehouse control server has pre-imported a precise Building Information Model (BIM) or 3D point cloud model of the warehouse. The model labels the number, location, and dimensions of all shelves, as well as the 3D coordinates (X, Y, Z) of the center point of each standard storage location (storage location), forming a storage location coordinate database. Simultaneously, the performance parameters of the deployed UAV1 are entered: maximum flight speed 2.5m / s, hovering accuracy ±0.1m, minimum obstacle avoidance distance 0.3m, and endurance 45 minutes (with scanning module). RFID scanning module parameters: dimensions 12cm x 8cm x 3cm, weight 180g, power consumption 7W, UHF band, and adjustable reading distance (for standard tags) 0.2-3 meters. A UWB positioning system is deployed within the warehouse, with base stations installed in the top corners of the warehouse, providing centimeter-level positioning for the UAVs across the entire area. Furthermore, the warehouse has full Wi-Fi 6 network coverage, providing a high-speed channel for data transmission.

[0090] Step S102: Review and plan tasks

[0091] The administrator initiates a "Full Area" task via a web client. Upon receiving the instruction, the warehouse control server starts the flight path planning engine.

[0092] Task Analysis: The inventory area is determined to be all storage locations in Zone A.

[0093] Path Generation: A "layered traversal" strategy is employed. First, a "backbone route" covering all main aisles in Area A is generated. This route is located in the center of the aisle and its height is adjustable. For each shelf level (e.g., 2-meter, 4-meter, ..., 12-meter levels), a horizontal route is generated that flies along the aisle at that height. The routes are in a continuous "S" or "bow" shape, ensuring that UAV 1 passes directly in front of every storage location. During planning, for the same level, the UAV flies along an aisle from beginning to end, scanning one side of the shelves; at the end of the aisle, it ascends / descends to a turning platform, then enters the adjacent aisle, scanning the other side of the shelves, and so on. This is called a "zigzag" or "serpentine" coverage strategy, resulting in a relatively short total path length.

[0094] Scan point and trigger zone binding: On the flight path, a "virtual scan point" (VSP) is set for each storage location that needs to be inventoried. This point is located in the space at a certain distance (e.g., 1.0 meter) directly in front of the storage location. Simultaneously, a spherical "trigger zone" with a radius of 0.8 meters is defined for each VSP. When the drone 1 enters this spherical area, the trigger condition is considered met. In this embodiment, approximately 50,000 VSPs are generated. The coordinates (X_vsp, Y_vsp, Z_vsp) of each VSP are calculated from the storage location center coordinates (X_loc, Y_loc, Z_loc) and a preset distance d. For example, for shelves arranged along the X-axis, the VSP coordinates might be (X_loc - d, Y_loc, Z_loc), where d = 1.0 meter, and the negative sign indicates that the point is directly in front of the storage location (alley side).

[0095] Waypoint Serialization: All waypoints that UAV 1 needs to pass through sequentially (including takeoff point, each VSP, turning point, and landing point) are arranged in order, and the attributes of each point are labeled (coordinates, expected speed, whether associated with a scan action, etc.), generating a standard flight path file (such as JSON format). The flight path file also contains metadata such as mission ID, version number, estimated total flight distance, and time. A simplified waypoint description might be: {"id": 1024, "type": "scan_point", "coords":[105.3, 45.7, 4.0], "velocity": 1.5, "associated_location_id": "A010203", "trigger_radius": 0.8}, indicating that waypoint 1024 is a scan point with coordinates (105.3, 45.7, 4.0), expected speed of 1.5 m / s, associated cargo location ID A010203, and trigger radius of 0.8 meters.

[0096] Step S103: Preparation and Mission Issuance of Drone 1

[0097] Select a standby drone 1. The warehouse control server sends the flight path file and mission ID to the flight control system of drone 1 via Wi-Fi. Simultaneously, ground crew or personnel in the automated drone nest ensure that drone 1's battery is fully charged (>90%) and that the RFID module is functioning correctly. Drone 1 loads the flight path file and performs a pre-flight self-check. The self-check includes sensor verification, communication link testing, motor and rotor checks, and RFID module ping testing. After all checks pass, drone 1 reports a "ready" status to the server.

[0098] Step S104: Autonomous Flight and Automatic Scanning

[0099] Drone 1 automatically takes off from the nest and flies towards the starting waypoint in Area A. After entering the predetermined route, it flies at a cruising speed of 1.5 m / s.

[0100] Obstacle Avoidance: The UAV 1's forward-looking LiDAR and binocular vision continuously scan the area in front and to the sides. When an obstacle not marked in the model (such as a temporarily placed pallet truck) is detected on the planned flight path, the flight control system initiates a local path replanning algorithm (such as Dynamic Window Method (DWA)) to smoothly avoid the obstacle while maintaining a safe distance, and then returns to the original flight path as quickly as possible. The obstacle avoidance process is real-time, and the detour path is inserted as a local waypoint into the original flight path sequence. The flight control system also sends an "obstacle avoidance event" notification to the server through the data transmission unit, along with the approximate location of the obstacle, for the background to record and analyze.

[0101] Trigger Scan: The trigger control unit (which can be a software module within the flight controller or an independent microcontroller) operates at a frequency of 100Hz. It acquires the real-time pose of UAV 1 (obtained from flight controller fusion IMU, visual odometry, and UWB positioning data) and quickly compares it with the VSP coordinates pre-stored in the flight path file. It calculates the Euclidean distance between the current position of UAV 1 and each VSP. Assuming that at the current moment, the distance between UAV 1 and VSP_001 is 0.75 meters, which is less than the trigger radius of 0.8 meters, and that no trigger record has been made for cargo location 001 before, the trigger control unit immediately sends a "start scanning" command to the RFID reader 12 via serial port or CAN bus. The command may include the cargo location ID (001) associated with the VSP. To improve calculation efficiency, the trigger control unit uses spatial indexing (such as grid method or quadtree / octree) to manage VSPs, and only performs distance calculations for VSPs within a specific range around the current UAV position.

[0102] RFID Reading: Upon receiving an instruction, the RFID reader 12 completes power-on, transmits an RF query signal, and receives and decodes the tag response within a very short time (e.g., 50ms). Because the gain of the directional antenna 13 is directed towards the shelf, it can effectively read one or more pallet tags affixed to location 001. The read tag EPC code and possible user data are encapsulated into a data packet. A successful read may obtain the EPCs of multiple tags. The reader records the signal strength (RSSI) and phase information of each tag; this information can be optionally uploaded to help determine the approximate location of the tag or filter out misreads from a distance.

[0103] Data transmission: The data packet is immediately sent via the Wi-Fi 6 network card on Drone 1 to the wireless access point (AP) deployed within the warehouse, and then transmitted to the warehouse control server via the wired network. The data packet contains the task ID, Drone 1 ID, timestamp, associated storage location ID (001), and a list of read tag information. The transmission protocol ensures the reliability and order of the data. If the network is temporarily unavailable, the data packet is stored in a local cache queue and retransmitted after the network is restored.

[0104] Step S105: Data reception and real-time processing

[0105] The data receiving service on the warehouse control server continuously listens on the network port. Upon receiving a data packet from drone 1:

[0106] Parsing and Verification: Parse the data packet and verify the validity of the task ID and drone ID.

[0107] Location-Label Binding: Based on the "Associated Location ID" (001) in the data packet, the read list of tags is bound to location 001 in the inventory database. The system records "At time T, UAV U scanned the tag set {L1, L2, ...} at location 001".

[0108] Deduplication and Filtering: Due to the multipath effect of RFID and the beam range of antenna 13, drone 1 may trigger multiple reads consecutively when approaching the storage location. The server sets a time window (e.g., data within 3 seconds for the same storage location ID) to deduplicatize duplicate tags (EPCs) and may filter out weak misread signals from adjacent storage locations based on signal strength index (RSSI). For example, only the record with the strongest RSSI for each EPC code within the time window is retained. Records with RSSI below a certain threshold may be directly considered as interference and ignored.

[0109] Inventory Update: The processed "location-tag" mapping is written to the temporary results table of this inventory count task. Optionally, the "Last Inventory Count Time" field in the inventory database is also updated. In scenarios with high real-time requirements, the inventory snapshot in the WMS can be updated immediately with the scan results, providing a progressive inventory view even if the inventory count task is not yet fully completed.

[0110] Step S106: Continuous Operations and Task Completion

[0111] Drone 1 continues flying, repeating steps S104 and S105 to scan all cargo locations in Area A one by one. The flight control system automatically controls Drone 1 to switch between different altitude layers (climb or descend) according to the flight path file. When Drone 1's battery level drops to a preset threshold (e.g., 25%) or when all scan points in the currently assigned area are completed, the flight control system controls it to return to the nest along the optimal path. Battery management is a crucial safety feature. The flight control system estimates the battery power required for return in real time and ensures that at all times, the remaining battery power is greater than the required return power plus a safety margin. If the estimation determines that the entire task cannot be completed, the flight control system will prematurely abort the mission and return to the nest, reporting the uncompleted portion to the server.

[0112] Step S107: Task Termination and Report Generation

[0113] After Drone 1 lands safely, the nest automatically connects its charging and data interfaces (such as USB) for rapid charging and complete uploading of flight logs and cached data. The warehouse control server confirms that all planned storage locations have been scanned (or marks locations that were not scanned due to malfunctions).

[0114] Final data comparison: The server runs a comparison algorithm to compare the "actual scanned inventory" in the temporary results table with the "book inventory" in the WMS system one by one.

[0115] Difference analysis: The following differences were identified:

[0116] Inventory shortage: The books show that there is stock, but the labels were not actually scanned.

[0117] Inventory surplus: No goods are recorded on the books, but the actual item was found with the label (possibly due to misplacement or failure to be promptly put into storage).

[0118] Quantity discrepancy: For storage locations that allow multiple goods to be stored together, the actual type or quantity of labels scanned may not match the records.

[0119] Misplacement: The scanned label belongs to goods that do not match the preset goods type of the storage location.

[0120] Report Generation: Automatically generates a detailed report including inventory overview (total number of storage locations, number of inventories counted, number of discrepancies), a detailed discrepancy list (location number, book information, actual information, discrepancy type), inventory time, and information on drones involved, and outputs it in PDF or Excel format. The report can be pushed to relevant personnel via email or system interface. The report can also include trend analysis, such as comparison with the previous inventory results, and statistics on frequently discrepant storage locations, providing deeper insights for management decisions.

[0121] Example 2: Multi-machine collaboration and dynamic partitioning

[0122] For ultra-large warehouses (such as those several times the size of Example 1), a single drone may not be able to complete the entire operation in one flight. This example demonstrates a multi-drone collaborative working mode.

[0123] In the route planning phase of step S102, the scheduling module of the warehouse control server dynamically divides the entire inventory area into three sub-areas (e.g., by aisle groups) based on the warehouse area division, the number of UAVs (assuming 3), and their respective endurance. An independent route file is generated for each sub-area, and the operation time is estimated to be approximately equal. The partitioning principle is load balancing and airspace isolation. Typically, partitioning is based on natural physical separation (e.g., different fire compartments, different warehouse areas) to ensure that the UAV operation areas do not overlap spatially, reducing coordination complexity.

[0124] In step S103, the three UAVs 1 are assigned tasks in sequence and the corresponding sub-area flight path files are issued.

[0125] Three drones (1) start operating simultaneously from their respective takeoff points (either sequentially from the same drone nest or from multiple drone nests distributed in different locations within the warehouse). The warehouse control server's monitoring screen displays the real-time location, status, progress, and battery level of the three drones (1).

[0126] The key to multi-aircraft collaboration lies in collision avoidance and airspace management. This embodiment adopts a "spatial + temporal" isolation strategy:

[0127] Spatial isolation: The planned sub-area flight paths do not overlap in physical space, ensuring that UAV 1 flies within its assigned lane group. This is the primary isolation method. During planning, it is ensured that the horizontal projection and vertical altitude layers of each UAV flight path are spatially staggered.

[0128] Time Synchronization and Communication: Each drone 1 maintains time synchronization with the server via a wireless network. Before entering areas where they may overlap (such as public take-off and landing zones or main roads), each drone must request "right of way" from the server. The server schedules drones according to the order of requests to avoid conflicts. For example, all drones share a take-off and landing platform located in the center of a warehouse. When a drone needs to land, it sends a landing request to the server. The server checks whether the platform is occupied and notifies other drones to postpone using the airspace until landing is complete.

[0129] Through multi-drone collaboration, the total inventory time can be reduced to approximately one-third of the time required for a single drone (considering scheduling overhead), significantly improving the timeliness of inventory checks in ultra-large warehouses. The server-side scheduling module also features dynamic load balancing. If one drone becomes idle due to a malfunction or early completion of its task, the scheduling module can dynamically allocate some unscanned storage locations from the task list of other drones still in operation to the idle drone, further optimizing the overall completion time.

[0130] Example 3: Adaptive Scanning and Retry Mechanism

[0131] To address the issue that some storage locations may fail to scan in one go due to damaged labels, poor orientation, or obstruction by metal, this embodiment introduces an adaptive scanning strategy.

[0132] In the trigger control unit logic of step S104, a feedback loop is added. After the RFID reader 12 completes a scan, it will return a "reading result status" to the trigger control unit, such as "reading successful (tag present)," "reading successful (empty storage location)," or "reading failed (no response)."

[0133] The trigger control unit determines the next action based on this state and the preset strategy:

[0134] If the message "Read successful (with label)" appears, record the success and continue flying.

[0135] If the message "Read successful (empty cargo space)" appears, the record is empty, and the flight continues.

[0136] If the read failure is "No response," the trigger control unit can execute a retry strategy. For example, it can immediately hover in place and instruct the RFID reader 12 to attempt scanning 1-2 more times at different power or frequency. If the retry is successful, it is recorded and continues; if it still fails, the location is marked as "scan failed" in the flight log, and the UAV's pose, RSSI, and other information may be recorded for subsequent manual review or adjustment of flight parameters. The retry strategy can be configured as a layered strategy: the first layer of retry involves slight pose adjustments (such as a 0.2-meter lateral shift or a 10-degree rotation); the second layer of retry involves increasing the transmission power; and the third layer of retry involves performing a small-radius circling scan. The number of retry attempts and parameters for each layer are configurable.

[0137] Furthermore, during route planning (S102), a lower flight speed or additional "hovering scan points" can be set for important cargo locations or cargo locations with low historical read rates to increase effective read opportunities. For example, for cargo locations storing high-value items, a "hovering scan" action can be set at its virtual scan point. After the drone arrives at the point, it hovers for 2 seconds, during which the RFID module performs multiple scans to improve the read success rate.

[0138] Example 4: Deep integration with WMS / WCS

[0139] This embodiment focuses on how the system of the present invention can be deeply integrated with existing warehouse management systems (WMS) and warehouse control systems (WCS) to form a complete closed loop of intelligent warehousing.

[0140] The warehouse control server provides standardized APIs (Application Programming Interfaces) or middleware interfaces to interact with the WMS.

[0141] Integration Point 1: Task Triggering. The WMS can automatically initiate inventory count task requests to the warehouse control server based on preset plans (such as every morning) or events (such as after a batch of goods has been received). For example, after completing the putaway operation of a batch of goods, the WMS automatically triggers a quick recount task for the storage locations involved in that batch to verify the accuracy of the putaway.

[0142] Integration Point 2: Data Synchronization. After completing the inventory count data comparison, the warehouse control server not only generates a report but also pushes confirmed inventory discrepancies (after review) directly to the WMS via API, driving the WMS to automatically generate inventory adjustment orders, transfer task orders, or trigger purchase alerts. This automates the process from inventory count discovery to system accounting correction, significantly shortening the business loop time.

[0143] Integration Point 3: Status Sharing. The warehouse control server of this system provides the real-time location and operational status of the drones to the WCS via an interface. When scheduling other automated equipment (such as AGVs and stacker cranes), the WCS can avoid the airspace where the drones are currently operating, achieving safe and collaborative operations across system devices. For example, when the WCS needs to schedule an AGV to enter a certain aisle to perform a pickup task, it will first query the system to confirm whether there are any drones operating in that aisle; if so, it will wait or choose an alternative path.

[0144] Integration Point 4: Model Synchronization. When the warehouse layout changes (such as adding, removing, or relocating shelves), the master data for storage locations in the WMS is updated and automatically synchronized to the warehouse 3D model and storage location coordinate database of this system via an interface, ensuring that inventory operations are based on the latest physical environment information. This maintains consistency between the digital model and the physical world, which is the foundation for the long-term reliable operation of the system.

[0145] Through the aforementioned deep integration, the system of this invention is no longer an independent inventory tool, but has been integrated into the very fabric of warehouse automation and informatization, becoming an indispensable "aerial sensing and data acquisition nerve" for smart warehouses.

[0146] Example 5: Optimization Strategies in Complex Environments

[0147] This embodiment illustrates the optimized application of the present invention in warehouses with dense metal shelving, complex electromagnetic environments, or special structures.

[0148] Challenge 1: Metal shelves strongly reflect and shield RFID signals. Optimization strategy:

[0149] Antenna optimization: Employ anti-metal tags or RFID tags optimized for metallic environments. Select directional antennas for the drone with narrower beamwidths and higher aspect ratios to reduce reflection interference from the metal shelf backplate.

[0150] Power and frequency tuning: Higher transmission power is configured for virtual scanning points in metal shelf areas during route planning. RFID readers support frequency hopping, automatically switching to clean frequencies when encountering interference.

[0151] Trigger strategy adjustment: For metal shelves, appropriately reduce the radius of the scanning trigger area and ensure that the virtual scanning point is closer to the storage location (e.g., 0.5 meters), so that the drone triggers the reading in the "sweet spot" area where the signal is strongest.

[0152] Challenge 2: The warehouse contains numerous obstacles such as pillars and pipes. Optimization strategy:

[0153] Refined modeling: All permanent obstacles are accurately marked in the 3D model of the warehouse, and set as no-fly zones during flight route planning for automatic detour.

[0154] Enhanced obstacle avoidance: Equip drones with more powerful 3D LiDAR or binocular vision systems to improve their ability to detect small obstacles (such as dangling cables).

[0155] Safe corridor: When planning flight routes, a "safe flight corridor" is defined in the center of the passageway. Drones are given priority to fly in this corridor, leaving sufficient margin to deal with obstacle avoidance maneuvers.

[0156] Challenge 3: Dynamic environment (e.g., occasional movement of people or forklifts). Optimization strategy:

[0157] Dynamic perception and response: Relying on onboard obstacle avoidance sensors to detect moving objects in real time. Once detected, it immediately slows down, hovers, or avoids the obstacle, and continues its mission only when it is safe to do so. Simultaneously, the drone can alert nearby personnel via light or sound signals.

[0158] Rule-making: Develop operational rules with warehouse management, such as restricting manual vehicle access to specific areas during drone inventory operations, or establishing clear operational warning zones.

[0159] Monitoring and intervention: Back-end monitoring personnel can view the drone's perspective at any time through the monitoring interface, and take over control through remote control commands when necessary to handle extreme situations.

[0160] The implementation principle of this invention is as follows: This invention discloses an automated inventory method and system for using drones equipped with RFID scanning in a warehouse, belonging to the field of warehousing and logistics automation technology. The method includes: a warehouse control server planning an inventory route for a drone 1, including scanning trigger points, based on a three-dimensional warehouse model; the drone 1 flying autonomously along the route, with the trigger control unit automatically activating the RFID scanning module to read tags when it approaches the trigger area of ​​the storage location; and the read data being transmitted back to the server in real time for binding and processing. This invention achieves full-process automation of inventory operations in automated warehouses, replacing inefficient and dangerous manual inventory checks. By utilizing the mobility of the drone 1 combined with non-contact RFID reading, it significantly improves inventory efficiency, safety, and accuracy. It has advantages such as flexible deployment, strong adaptability, and support for multi-drone collaboration, making it suitable for inventory management in various large-scale automated warehouses.

[0161] The core innovation of this invention lies in the systematic integration of the spatial maneuverability of drones, the non-contact identification capabilities of RFID, and intelligent path planning and trigger control technology based on precise 3D models, creatively proposing an automated inventory paradigm of "scanning as it flies." Through software-defined methods, it transforms fixed warehouse space into a dynamic workspace where drones can automatically perform data collection tasks, achieving a fundamental shift in inventory operations from "people finding goods" to "data flow proactively covering cargo locations." The system's modular design makes it easy to deploy and expand, while its deep integration with upper-level information systems ensures its value can be seamlessly integrated into the enterprise's overall logistics management system.

[0162] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An automated inventory counting method using RFID scanning via drones in a warehouse, characterized in that, Includes the following steps: Step S1: The warehouse control server generates the inventory flight path of the UAV (1) based on the three-dimensional spatial information of the target inventory area. The flight path includes multiple scanning trigger points associated with the inventory location and their corresponding scanning trigger area information. Step S2: The UAV (1) loads the inventory flight route, takes off from the starting point and flies autonomously along the route, while using the onboard obstacle avoidance sensor to perform real-time environmental perception and dynamic obstacle avoidance; Step S3: During flight, the trigger control unit continuously acquires the real-time location information of the UAV (1) and compares it with the information of the scanning trigger area; when it is determined that the UAV (1) enters the scanning trigger area of ​​any cargo location, the RFID scanning module on the UAV (1) is automatically triggered to read the RFID signal of the cargo location; Step S4: The RFID scanning module will send the RFID tag data it reads to the warehouse control server through the wireless data transmission unit; Step S5: The warehouse control server receives the RFID tag data, binds and processes it with the corresponding storage location information, and updates the inventory data; Step S6: After the UAV (1) completes the flight route, it automatically returns and lands.

2. The automated inventory method for warehouse use of RFID scanning by drones in accordance with claim 1, characterized in that, In step S1, generating the inventory flight route specifically includes: Based on the location, height, and coordinates of the shelves in the three-dimensional digital model of the warehouse, and combined with the flight performance parameters and safe obstacle avoidance distance of the UAV (1), a horizontal continuous flight path is generated for the height of each shelf that needs to be inventoried. On the path, a virtual scanning point is set at a preset distance in front of each inventory location. The virtual scanning point is associated with a three-dimensional scanning trigger area.

3. The automated inventory method for warehouse use of RFID scanning by drones in accordance with claim 2, characterized in that, The scan triggering area is a spherical area, a cubic area, or a fan-shaped area defined according to the direction of the shelf opening, with the virtual scan point as the center and a preset length as the radius.

4. The automated inventory method for warehouse use of RFID scanning by drones in accordance with claim 1, characterized in that, In step S3, the step of automatically triggering the RFID scanning module to read the data further includes: After triggering the read, it receives read feedback from the RFID scanning module; If the reading feedback is a failure, then according to the preset retry strategy, the drone (1) is controlled to perform at least one of the following operations: hovering, adjusting pose, or adjusting RFID reading parameters, and then the reading is triggered again.

5. The automated inventory method for warehouse use of RFID scanning by drones in accordance with claim 1, characterized in that, The RFID scanning module includes a miniaturized UHF RFID reader (12) and a directional antenna (13), the main radiation direction of which is directed toward the shelf on one side of the flight path; The trigger control unit is integrated into the flight controller of the UAV (1) or connected to the flight controller as an independent module via a communication interface.

6. An automated inventory system for warehouses, using drones equipped with RFID scanning, characterized in that, An automated inventory management method for using an RFID scanner mounted on a drone within a warehouse, as described in any one of claims 1-5, comprises: The warehouse control server is equipped with a 3D warehouse model, a cargo location coordinate database, and a flight path planning module. It is used to generate inventory flight paths and receive and process inventory data. At least one unmanned aerial vehicle (1), including the unmanned aerial vehicle body (11), flight control system, positioning module, obstacle avoidance sensor and power supply unit, is used to fly autonomously according to the inventory flight route; An RFID scanning module, fixed on the UAV body (11), includes an RFID reader (12) and an antenna (13), used to read RFID tag information on the cargo location after triggering; The trigger control unit is connected in communication with the flight control system of the UAV (1) and the RFID scanning module, and is used to automatically control the start and stop of the RFID scanning module according to the matching result of the real-time position of the UAV (1) and the preset scanning trigger area; A data transmission unit is installed on the UAV (1) and is used to establish a communication link between the UAV (1) and the warehouse control server to transmit flight route instructions and inventory data.

7. An automated inventory system for warehouses using RFID scanning via drones, as described in claim 6, is characterized in that... The route planning module of the warehouse control server is specifically used to: generate a route file containing a series of ordered waypoints based on the warehouse 3D model, the performance constraints of the UAV (1) and the inventory task range, wherein an associated virtual scanning point and a 3D trigger area centered on each inventory location are generated; the route file is sent to the UAV (1) through the data transmission unit.

8. An automated inventory system for warehouses using RFID scanning via drones, as described in claim 6, is characterized in that... The trigger control unit is configured to: acquire real-time pose data of the UAV (1) provided by the flight control system at a fixed frequency, calculate the spatial distance between the UAV and each virtual scanning point in the flight path file, and generate a trigger command to be sent to the RFID scanning module when the distance is less than a preset trigger threshold.

9. An automated inventory system for warehouses using RFID scanning via drones, as described in claim 6, is characterized in that... The system also includes at least one UAV (1) nest, which is equipped with a charging interface and a data interface for automatically charging the UAV (1) after landing and for retransmitting cached data via wired means; the warehouse control server also includes a multi-machine scheduling module for performing task partitioning, route allocation and conflict coordination when there are multiple UAVs (1).

10. An automated inventory system for warehouses using RFID scanning via drones, as described in claim 6, characterized in that... The warehouse control server also includes a data receiving and processing module, which is used to parse, verify, deduplicate, and bind the received RFID tag data with the storage location information. It further includes a report generation module, which is used to compare the processed actual inventory data with the system's book inventory after the inventory task is completed, identify differences, and generate an inventory report.

Images