Search

CN-121977547-A - Shared tray self-organizing positioning and map building method and system based on distributed crowd sensing

CN121977547ACN 121977547 ACN121977547 ACN 121977547ACN-121977547-A

Abstract

The invention discloses a shared tray self-organizing positioning and map constructing method and system based on distributed crowd sensing, wherein the method comprises the steps of collecting self-motion data through an inertial measurement unit built in each tray, and exchanging motion data and signal intensity information with an adjacent tray in a detection range by utilizing a near field communication module; the method comprises the steps of exchanging motion data and signal intensity information, forming a local relative pose network according to the exchanged motion data and the exchanged signal intensity information, combining the local relative pose network, iteratively calculating global relative coordinates of all trays through a cluster consistency optimization algorithm, synchronously deducing geometric boundaries of fixed obstacles and passable areas in a station, and outputting a real-time dynamic tray cluster distribution map and a station structure map according to the global relative coordinates and the geometric boundaries. By utilizing the embodiment of the invention, the high-precision self-organizing positioning and map construction of the shared tray cluster under the condition of no external infrastructure can be realized, and the flexibility, adaptability and deployment efficiency of the system are improved.

Inventors

  • XU ZHIHAO
  • WANG KAI

Assignees

  • 浙江久鼎智联科技有限公司

Dates

Publication Date
20260505
Application Date
20260407

Claims (10)

  1. 1. A shared tray self-organizing positioning and map construction method based on distributed crowd sensing is characterized by comprising the following steps: Acquiring self motion data through an inertial measurement unit built in each tray, and exchanging motion data and signal intensity information with an adjacent tray in a detection range by utilizing a near field communication module; According to the exchanged motion data and signal intensity information, based on a multi-tray collaborative geometry solving method, the relative distance and heading relation between any two trays are estimated in a distributed mode, and a local relative pose network is formed; Combining the local relative pose network, iteratively calculating global relative coordinates of each tray through a cluster consistency optimization algorithm, and synchronously deducing the geometric boundaries of the fixed obstacles and the passable area in the station; And carrying out on-line compensation and correction on motion estimation of the inertial measurement unit according to the global relative coordinates and the geometric boundary, and outputting a real-time dynamic tray cluster distribution map and a station structure map.
  2. 2. The method according to claim 1, wherein the acquiring self motion data by the inertial measurement unit built in each tray and exchanging motion data and signal strength information with the adjacent tray within the detection range using the near field communication module comprises: acquiring triaxial acceleration and triaxial angular velocity data in real time through an inertial measurement unit arranged in a tray, and carrying out noise reduction and drift compensation processing on the original data by adopting a Kalman filtering algorithm to generate preprocessed motion data; Dividing the preprocessed motion data according to time windows, extracting linear acceleration mean value, angular velocity variance and displacement integral characteristics in each time window, and generating motion characteristic vectors; periodically broadcasting a data packet containing a unique identifier of the tray and a motion characteristic vector through a near field communication module, and simultaneously scanning and receiving the same type of data packet broadcasted by the adjacent tray to generate an adjacent tray data exchange record; And analyzing signal intensity indication values in the data exchange records of the adjacent trays, and constructing a local adjacent information table by combining the motion characteristic vectors and the time stamp information.
  3. 3. The method according to claim 2, wherein the step of distributively estimating the relative distance and heading relation between any two trays based on a multi-tray collaborative geometry solving method according to the exchanged motion data and signal intensity information to form a local relative pose network comprises: Extracting a signal strength indication value sequence between each adjacent tray from a local adjacent information table, converting a signal strength value into a distance estimation value by adopting a path loss model, and generating an initial distance estimation matrix; Combining the motion feature vector of the self and the motion feature vector of the adjacent tray, and performing iterative correction on the initial distance estimation matrix by using an extended Kalman filter through calculating the correlation of the motion trail to generate a corrected distance matrix; Based on the corrected distance matrix, calculating the relative position coordinates of each tray in a local coordinate system by adopting a multidimensional scale positioning algorithm, and calculating a relative course angle by utilizing gyroscope data fusion to generate a local relative pose map; And carrying out consistency check on the local relative pose graph, and eliminating measurement errors through least square method optimization to form a stable local relative pose network.
  4. 4. A method according to claim 3, wherein the iterative calculation of global relative coordinates of each pallet by means of a cluster consistency optimization algorithm in combination with the local relative pose network synchronously infers geometric boundaries of fixed obstacles and passable areas in the station, comprising: Carrying out graph theory modeling on nodes and edges in the local relative pose network, and adopting a distributed graph optimization algorithm to parallelly calculate a transformation matrix from a local coordinate system to a global coordinate system on each tray node to generate initial global coordinate estimation; Aligning tray coordinates of the overlapped area through an iterative nearest point algorithm, optimizing continuity of initial global coordinates by utilizing loop consistency constraint, and generating an optimized global relative coordinate set; based on the optimized global relative coordinate set, analyzing abnormal stagnation points and steering modes in the tray movement track, identifying potential fixed obstacle positions, and generating an obstacle assumed position set; And performing density cluster analysis on the assumed position set of the obstacle, fitting out the geometric boundary outline of the obstacle, and generating a passable area boundary according to the normal passing track of the tray to form a station structure geometric boundary.
  5. 5. The method of claim 4, wherein the online compensation and correction of motion estimation of the inertial measurement unit based on the global relative coordinates and the geometric boundary outputs a real-time dynamic pallet cluster map and a station structure map, comprising: comparing the displacement estimated by the inertial measurement unit with the global relative coordinate, calculating the accumulated error of the inertial measurement unit, and generating an error compensation model by adopting an adaptive filter; Performing online correction on the inertial measurement unit data acquired in real time by using an error compensation model, and performing motion track optimization by combining geometric boundary constraint to generate a corrected real-time position and posture of the tray; fusing all corrected real-time positions and postures of the trays, and generating a tray cluster distribution map under the same time reference system through time synchronization and spatial interpolation; and stacking the pallet cluster distribution map and the geometric boundary of the station structure, maintaining the real-time performance of the map by adopting an incremental map updating mechanism, and outputting a complete station structure map.
  6. 6. A shared tray self-organizing positioning and map building system based on distributed crowd sensing, the system comprising: The acquisition module is used for acquiring motion data of the acquisition module through an inertial measurement unit built in each tray and exchanging motion data and signal intensity information with an adjacent tray in a detection range by utilizing the near field communication module; The estimating module is used for estimating the relative distance and course relation between any two trays in a distributed manner based on a multi-tray collaborative geometry solving method according to the exchanged motion data and signal intensity information to form a local relative pose network; The inference module is used for combining the local relative pose network, iteratively calculating global relative coordinates of each tray through a cluster consistency optimization algorithm, and synchronously inferring the geometric boundaries of the fixed obstacles and the passable area in the station; And the output module is used for carrying out on-line compensation and correction on the motion estimation of the inertial measurement unit according to the global relative coordinates and the geometric boundary and outputting a real-time dynamic tray cluster distribution map and a station structure map.
  7. 7. The system according to claim 6, wherein the acquisition module is specifically configured to: acquiring triaxial acceleration and triaxial angular velocity data in real time through an inertial measurement unit arranged in a tray, and carrying out noise reduction and drift compensation processing on the original data by adopting a Kalman filtering algorithm to generate preprocessed motion data; Dividing the preprocessed motion data according to time windows, extracting linear acceleration mean value, angular velocity variance and displacement integral characteristics in each time window, and generating motion characteristic vectors; periodically broadcasting a data packet containing a unique identifier of the tray and a motion characteristic vector through a near field communication module, and simultaneously scanning and receiving the same type of data packet broadcasted by the adjacent tray to generate an adjacent tray data exchange record; And analyzing signal intensity indication values in the data exchange records of the adjacent trays, and constructing a local adjacent information table by combining the motion characteristic vectors and the time stamp information.
  8. 8. The system according to claim 7, wherein the estimation module is specifically configured to: Extracting a signal strength indication value sequence between each adjacent tray from a local adjacent information table, converting a signal strength value into a distance estimation value by adopting a path loss model, and generating an initial distance estimation matrix; Combining the motion feature vector of the self and the motion feature vector of the adjacent tray, and performing iterative correction on the initial distance estimation matrix by using an extended Kalman filter through calculating the correlation of the motion trail to generate a corrected distance matrix; Based on the corrected distance matrix, calculating the relative position coordinates of each tray in a local coordinate system by adopting a multidimensional scale positioning algorithm, and calculating a relative course angle by utilizing gyroscope data fusion to generate a local relative pose map; And carrying out consistency check on the local relative pose graph, and eliminating measurement errors through least square method optimization to form a stable local relative pose network.
  9. 9. A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of claims 1-5 when run.
  10. 10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, the processor being arranged to run the computer program to perform the method of any of claims 1-5.

Description

Shared tray self-organizing positioning and map building method and system based on distributed crowd sensing Technical Field The invention belongs to the technical field of trays, and particularly relates to a shared tray self-organizing positioning and map building method and system based on distributed crowd sensing. Background In scenes such as warehouse logistics and intelligent factories, the trays are used as core cargo bearing units, and accurate positioning and real-time map construction of the environment are important for realizing efficient automation. Existing solutions rely primarily on pre-deployed fixed infrastructure such as ultra wideband base stations or visual identification. The method has the problems of high deployment cost, poor flexibility, difficulty in adapting to dynamic change environment and the like. In addition, a single tray is positioned only by a sensor of the tray, so that accumulated errors are easy to generate and the tray is easy to fail in a complex shielding environment. Although the multi-agent cooperative positioning technology has been explored in the robot field, the multi-agent cooperative positioning technology is applied to a shared tray cluster with limited resources and variable scale, and realizes self-organizing map construction without an external infrastructure, and still faces a plurality of technical challenges such as distributed computation, data fusion, system robustness and the like. Disclosure of Invention The invention aims to provide a shared tray self-organizing positioning and map building method and system based on distributed crowd sensing, which are used for solving the defects in the prior art, realizing high-precision self-organizing positioning and map building of a shared tray cluster under the condition of no external infrastructure, and improving the flexibility, adaptability and deployment efficiency of the system. The embodiment of the application provides a shared tray self-organizing positioning and map building method based on distributed crowd sensing, which comprises the following steps: Acquiring self motion data through an inertial measurement unit built in each tray, and exchanging motion data and signal intensity information with an adjacent tray in a detection range by utilizing a near field communication module; According to the exchanged motion data and signal intensity information, based on a multi-tray collaborative geometry solving method, the relative distance and heading relation between any two trays are estimated in a distributed mode, and a local relative pose network is formed; Combining the local relative pose network, iteratively calculating global relative coordinates of each tray through a cluster consistency optimization algorithm, and synchronously deducing the geometric boundaries of the fixed obstacles and the passable area in the station; And carrying out on-line compensation and correction on motion estimation of the inertial measurement unit according to the global relative coordinates and the geometric boundary, and outputting a real-time dynamic tray cluster distribution map and a station structure map. Still another embodiment of the present application provides a system for self-organizing and mapping a shared tray based on distributed crowd sensing, the system comprising: The acquisition module is used for acquiring motion data of the acquisition module through an inertial measurement unit built in each tray and exchanging motion data and signal intensity information with an adjacent tray in a detection range by utilizing the near field communication module; The estimating module is used for estimating the relative distance and course relation between any two trays in a distributed manner based on a multi-tray collaborative geometry solving method according to the exchanged motion data and signal intensity information to form a local relative pose network; The inference module is used for combining the local relative pose network, iteratively calculating global relative coordinates of each tray through a cluster consistency optimization algorithm, and synchronously inferring the geometric boundaries of the fixed obstacles and the passable area in the station; And the output module is used for carrying out on-line compensation and correction on the motion estimation of the inertial measurement unit according to the global relative coordinates and the geometric boundary and outputting a real-time dynamic tray cluster distribution map and a station structure map. A further embodiment of the application provides a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of the preceding claims when run. Yet another embodiment of the application provides an electronic device comprising a memory having a computer program stored therein and a processor configured to run the computer program to perform the method recited in any of the