AGV adaptive navigation method based on multi-sensor fusion

Abstract

This invention discloses an AGV adaptive navigation method based on multi-sensor fusion, relating to the field of AGV navigation technology. The method includes: calculating unified timing and reference heading for the AGV, and converting the AGV's navigation task into a ternary command skeleton; determining the effectiveness of BeiDou navigation and generating the effectiveness level of the AGV's current reference heading; constructing passable sectors and a navigation constraint field; generating shared passable area driving instructions based on entry time windows; and performing heading correction and time window constraint superposition on the ternary command skeleton under a unified timing time axis based on the navigation constraint field and the shared passable area driving instructions, continuously expanding and updating it to generate an AGV navigation command stream. This invention achieves a unified approach to AGV heading stability, time-controllable navigation, and adaptive navigation by continuously expanding and updating the ternary command skeleton to form a real-time updatable navigation command stream.

Description

Technical Field

[0001] This invention relates to the field of AGV navigation technology, and in particular to an AGV adaptive navigation method based on multi-sensor fusion. Background Technology

[0002] High-precision positioning and timing using BeiDou, dual-antenna orientation, inertial measurement unit (IMU) heading estimation, and local obstacle avoidance using lidar / ultrasound are common key technologies in AGV and mobile robot navigation systems. In engineering, a fusion framework of "GNSS / BeiDou positioning + IMU inertial navigation + lidar or vision" is typically adopted. State estimation is achieved through filters, and feasible speed and heading commands are generated based on methods such as local cost maps or dynamic windows. In multi-vehicle collaborative scenarios, area mutual exclusion, intersection yielding, or scheduling tables are often combined to manage lane occupancy, reducing the probability of conflicts in shared areas such as narrow passages and intersections.

[0003] However, most current methods rely on continuous positioning quality, map-based positioning consistency, and historical integration stability. Once GNSS heading calculation degrades or multiple source timelines are out of sync, fusion estimation is prone to timing deviations and heading drift, which in turn leads to control command jitter and unstable path following. Meanwhile, obstacle avoidance and passage reservation are mostly driven by global scheduling or strong map semantic constraints, which makes navigation unstable under weak mapping / no map conditions and makes it difficult to control the passage timing in shared areas. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an AGV adaptive navigation method based on multi-sensor fusion to solve the problems of unstable navigation and difficulty in controlling the passage timing in shared areas under weak mapping / no map conditions.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an AGV adaptive navigation method based on multi-sensor fusion, comprising,

[0008] The Beidou receiver and dual-antenna structure are used to calculate the unified timing and reference heading of the AGV, and the navigation task of the AGV is converted into a three-element command skeleton.

[0009] Based on the three-dimensional command skeleton, the inertial measurement unit is used to determine the validity of the reference heading and generate the validity level of the AGV's current reference heading;

[0010] By using lidar and ultrasonic sensors to determine the nearest obstacle distance in the AGV's forward direction area, a passable sector is constructed, and this is fused with the effectiveness level of the AGV's current reference heading to obtain a navigation constraint field.

[0011] Based on the navigation constraint field, before the AGV enters the preset shared passage area, the entry time window is calculated according to the current speed of the AGV, and the entry time window is reserved under unified timing to generate the driving instruction for the shared passage area;

[0012] Based on the navigation constraint field and the driving instructions of the shared passage area, the ternary instruction skeleton is subject to heading correction and time window constraint superposition under a unified time axis, and is continuously expanded and updated to generate AGV navigation instruction flow.

[0013] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for converting the AGV's navigation task into a ternary instruction skeleton are as follows:

[0014] The AGV's navigation task is broken down into a sequence of target stations and a desired timeline;

[0015] Calculate the next target direction angle based on the spatial coordinate difference between the AGV's current position and the next target station in the target station sequence;

[0016] The next target direction angle, the expected timetable, and the AGV's reference heading are combined and encapsulated according to the target station sequence to obtain the ternary instruction skeleton.

[0017] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for determining the validity of the reference heading are as follows:

[0018] The navigation control cycle is set during the AGV's travel between adjacent target stations;

[0019] The angular velocity of the AGV's geometric center is measured by an inertial measurement unit, and the angular velocity is integrated within the navigation control cycle to obtain the AGV's heading change. At the same time, the AGV's reference heading change is calculated.

[0020] During the navigation control cycle, the difference between the change in heading and the change in the AGV's reference heading is calculated to obtain the heading difference.

[0021] A heading difference threshold is set based on the deviation between the heading change and the AGV reference heading.

[0022] If the heading difference exceeds the heading difference threshold in multiple consecutive navigation control cycles, the validity level of the AGV's current reference heading is determined to be invalid.

[0023] Conversely, if the heading difference does not exceed the heading difference threshold in multiple consecutive navigation control cycles, the validity level of the AGV's current reference heading is determined to be valid.

[0024] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for constructing the passable sector are as follows:

[0025] Based on the spatial point cloud information output by the lidar, calculate the nearest distance to obstacles in the direction the AGV is moving, and divide the passable sectors of the AGV.

[0026] Ultrasonic sensors are used to detect obstacles in the nearby area around the AGV, and areas with obstacles are marked as prohibited sectors. Prohibited sectors are then removed from the passable sectors.

[0027] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the traversable sector is fused with the effectiveness level of the AGV's current reference heading. The specific steps are as follows:

[0028] When the validity level of the current reference heading of the AGV is valid, the AGV travels within the passable sector according to the reference heading;

[0029] When the validity level of the current reference heading of the AGV is invalid, no directional constraint on the reference heading is applied to the passable sector.

[0030] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for reserving an entry time window under unified timing are as follows:

[0031] The entry time window is compared with the time window occupancy table of the shared passage area. By judging the non-overlapping time intervals, the entry time windows that do not overlap with the entry time windows of other AGVs are selected.

[0032] During the selection process for entering the time window, the AGV's start-up delay time and braking response time are added to the entry time window as fixed safety time conditions.

[0033] When there are entry time windows in the time window occupancy table that meet the filtering criteria, the entry time window with the earliest start time of the appointment is selected.

[0034] When there is no entry time window in the time window occupancy table that meets the filtering criteria, the AGV moves to a safe position to wait, and at the same time, applies for the next earliest entry time window.

[0035] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for performing heading correction and time window constraint superposition on the ternary command skeleton are as follows:

[0036] The ternary command skeleton is modified using a navigation constraint field;

[0037] The shared passage area driving instruction is inserted into the ternary instruction skeleton, and timing boundary constraints are applied to the ternary instruction skeleton.

[0038] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for correcting the ternary command skeleton using a navigation constraint field are as follows:

[0039] Map the next target heading angle in the ternary command skeleton to the heading range covered by the traversable sector;

[0040] When the next target direction angle in the ternary command skeleton enters the heading range covered by the passable sector, the next target direction angle in the ternary command skeleton remains unchanged.

[0041] When the next target heading angle in the ternary command skeleton does not fall within the heading range covered by the passable sector, select the heading direction with the smallest deviation from the next target heading angle in the ternary command skeleton from the passable sector, and replace the next target heading angle in the ternary command skeleton with the selected heading direction.

[0042] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for applying temporal boundary constraints to the ternary instruction skeleton are as follows:

[0043] When the shared passage area driving instruction is empty, the three-dimensional instruction skeleton only retains the requirements of the expected timetable;

[0044] When the shared passage area driving instruction is not empty, the entry time and exit deadline of the shared passage area driving instruction are superimposed on the three-element instruction skeleton.

[0045] As a preferred embodiment of the AGV adaptive navigation method based on multi-sensor fusion described in this invention, the specific steps for generating the AGV navigation command stream are as follows:

[0046] After completing the course correction and time window constraint superposition, the next target direction angle, expected time table, and entry and exit times of the shared passage area at each consecutive moment are combined into a navigation command under unified timing, and continuously output in the order of the unified timing time axis to obtain the navigation command flow of the AGV.

[0047] The beneficial effects of this invention are as follows: It utilizes BeiDou second pulses to establish unified timing, aligning perception, control, and multi-vehicle time window occupancy tables on the same time axis, reducing the risk of timing drift in multi-source data fusion and command issuance. Dual-antenna calculation yields an absolute reference heading consistent with true north. It conditionally fuses the passable sectors constructed by lidar and ultrasound with the heading effectiveness level, enhancing long-term heading stability when the heading is effective, and degenerating into purely local safety decisions when the heading is ineffective, ensuring continuity and safety. Before entering a shared passage area, it reserves entry time windows based on unified timing and generates window commands, lightweightly coupling local passage with multi-vehicle timing mutual exclusion, reducing conflicts without introducing heavy global scheduling complexity. Finally, through continuous expansion and updating of the ternary command skeleton, it forms a real-time updatable navigation command stream, achieving a unified AGV heading stability, timing controllability, and adaptive navigation. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of an AGV adaptive navigation method based on multi-sensor fusion.

[0050] Figure 2 This is a flowchart of the ternary instruction skeleton conversion process.

[0051] Figure 3 This is a flowchart for determining the effectiveness of BeiDou navigation.

[0052] Figure 4 Flowchart for reserving entry time windows for shared access areas. Detailed Implementation

[0053] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0054] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0055] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0056] Reference Figures 1-4 This is one embodiment of the present invention, which provides an AGV adaptive navigation method based on multi-sensor fusion, including the following steps:

[0057] S1. The Beidou receiver and dual-antenna structure are used to calculate the unified timing and reference heading of the AGV, and the navigation task of the AGV is converted into a three-element command skeleton.

[0058] By outputting a second pulse signal synchronized with the BeiDou navigation satellites through the BeiDou receiver, the second pulse signal is used as the time alignment reference for the AGV, so that the internal running time of the AGV is uniformly constrained on the same BeiDou time axis, thus completing the unified time synchronization of the AGV and providing a unique time reference for subsequent adaptive navigation of the AGV.

[0059] Simultaneously, the reference heading of the AGV is obtained by processing the satellite signals from the BeiDou navigation satellites through a dual-antenna structure. Specifically, two BeiDou antennas (i.e., a dual-antenna structure) are set on the AGV along its direction of travel. Under unified timing conditions, the received satellite signals are processed independently through the dual-antenna structure to obtain the spatial positions of the two BeiDou antennas. The coordinate difference between the spatial positions of the two BeiDou antennas is calculated to obtain the direction vector of the BeiDou antenna in the geographic coordinate system, which is the orientation direction vector of the AGV in the geographic coordinate system. By calculating the angle between the orientation direction vector of the AGV and the geographic true north direction vector, the reference heading of the AGV is obtained.

[0060] The dual-antenna structure receives satellite signals simultaneously transmitted by multiple BeiDou navigation satellites. These signals include time information (based on the same unified time synchronization standard) and satellite orbital positions. The time information is used to obtain the propagation time of the satellite signal from the satellite to the BeiDou antenna. Based on the propagation speed and time of the satellite signal, the spatial distance between the BeiDou navigation satellite and the BeiDou antenna is estimated. Using the satellite orbital positions, the estimated spatial distances between the multiple BeiDou navigation satellites and the BeiDou antenna are differentially processed to obtain the spatial position of the BeiDou antenna.

[0061] The AGV's navigation task is broken down into a sequence of target stations and a desired timetable, and then combined with the AGV's current reference heading to obtain a three-dimensional instruction skeleton. The specific steps are as follows:

[0062] The target station sequence refers to the order in which the AGV travels, and the expected timetable refers to the time interval during which the AGV is allowed to arrive at each station under unified time synchronization.

[0063] Calculate the spatial coordinate difference between the AGV's current position and the next target station, and obtain the target azimuth angle (the actual travel direction from the current position to the next target station) by taking the arctangent of the horizontal coordinate difference. Calculate the difference between the AGV's current reference heading and the target azimuth angle, and normalize it to the interval (−π, π] to obtain the next target heading angle. Combine and encapsulate the next target heading angle, the expected timetable, and the AGV's reference heading according to the target station sequence to obtain the ternary instruction skeleton.

[0064] It should be noted that by utilizing BeiDou time synchronization and a dual-antenna structure to jointly construct a stable navigation reference consistent with true geographical north, the AGV can obtain an absolutely meaningful time reference and heading reference without relying on historical motion accumulation, environmental mapping, or the fusion of positioning results. This provides a unified, reliable, and drift-free foundation for subsequent multi-sensor fusion, adaptive navigation decision-making, and heading stability during long-term operation.

[0065] S2. Based on the three-dimensional command skeleton, the inertial measurement unit is used to determine the validity of the reference heading and generate the validity level of the AGV's current reference heading.

[0066] During the journey between adjacent target stations, the AGV periodically performs navigation judgments and control updates. Specifically, based on the AGV's current reference heading and unified timing, the three-dimensional instruction skeleton is updated. Each navigation judgment and control update constitutes a navigation control cycle. Each navigation control cycle is much shorter than the journey time between adjacent target stations. For example, if the journey time between two adjacent target stations is 10 minutes, the navigation control cycle can be set to 10 milliseconds.

[0067] During each navigation control cycle, the angular velocity of the AGV's geometric center is measured by the inertial measurement unit, and the angular velocity is integrated during the navigation control cycle to obtain the AGV's heading change, which characterizes the AGV's attitude change. The heading change is then time-aligned under unified timing.

[0068] Within a navigation control cycle, the difference between the change in heading and the change in the AGV's reference heading is calculated to obtain the heading difference value. A preset heading difference threshold is used, and the heading difference value is continuously compared with the heading difference threshold over multiple navigation control cycles. If the heading difference value exceeds the heading difference threshold for multiple consecutive navigation control cycles (e.g., 3 cycles to avoid situations where instantaneous noise and brief obstructions cause the validity level of the AGV's current reference heading to be invalid), it is determined that there is a continuous inconsistency between the AGV's reference heading and the AGV's actual heading, indicating that the validity level of the AGV's current reference heading is invalid. Conversely, if the heading difference value does not exceed the heading difference threshold for multiple consecutive navigation control cycles, it is determined that there is a continuous consistency between the AGV's reference heading and the AGV's actual heading, indicating that the validity level of the AGV's current reference heading is valid.

[0069] In this process, under the condition that the BeiDou navigation system is in a real-time dynamic positioning fixed solution state and the AGV is traveling at low or constant speed in a straight line, the natural fluctuation range of the AGV's reference heading within the navigation control cycle is recorded as a reference range for normal changes in the AGV's reference heading. The heading changes are repeatedly collected under the same conditions (the AGV traveling at low or constant speed in a straight line within the navigation control cycle), and the difference between the extreme values ​​of the repeatedly collected heading changes and the change in the AGV's reference heading is taken as the reasonable deviation range of the heading changes within the navigation control cycle. The upper limit of the reasonable deviation range between the change in the AGV's reference heading and the actual heading changes is set as the heading difference threshold. For example, if the natural fluctuation range of the AGV's reference heading is Q degrees within the same navigation control cycle, and the maximum value of the repeatedly collected heading changes is P degrees and the minimum value is I degrees, then the reasonable deviation range between the extreme values ​​of the heading changes and the actual heading changes is... That is, the heading difference threshold is .

[0070] It should be noted that by using BeiDou navigation to provide unified timing for AGVs and obtain absolute heading references, a stable and consistent spatiotemporal reference is established for AGVs during navigation tasks. This provides reliable time consistency and heading references for subsequent collaborative processing of multi-sensor information and navigation control, avoiding the problem of unstable navigation decisions caused by time asynchrony or heading reference drift.

[0071] S3. Using lidar and ultrasonic sensors, determine the nearest obstacle distance in the AGV's forward direction area, construct a passable sector, and fuse it with the effectiveness level of the AGV's current reference heading to obtain a navigation constraint field.

[0072] Based on the unified timing of the BeiDou navigation satellites, the real-time outputs of the lidar and ultrasonic sensors are synchronously connected.

[0073] In the local coordinate system of the AGV, based on the spatial point cloud information output by the LiDAR, the nearest distance to obstacles in the AGV's forward direction is calculated, and the AGV's passable sectors are divided. Specifically, the LiDAR acquires spatial point cloud information surrounding the AGV and converts the point cloud coordinates into a local coordinate system with the AGV's geometric center as the origin and the AGV's forward direction as the reference axis. The space in the AGV's forward direction is divided along the left and right sides, resulting in multiple adjacent directional angle intervals. Within each directional angle interval, the LiDAR compares the distances of all point cloud data and selects the point cloud distance with the smallest distance from the AGV's geometric center as the nearest obstacle distance. Based on the AGV's current operating speed and the minimum safe braking distance at the current speed, directional angle intervals that meet the requirement that the nearest obstacle distance is greater than the minimum safe braking distance are marked as passable directional angle intervals, while directional angle intervals that do not meet the requirement that the nearest obstacle distance is greater than the minimum safe braking distance are marked as impassable directional angle intervals. This completes the construction of passable sectors within the AGV coordinate system.

[0074] The minimum safe braking distance can be calculated in two parts. The first part is the distance traveled at the speed before braking under the braking response time, which is the product of the speed before braking and the braking response time. The second part is the AGV uniform deceleration distance, which is calculated by using the uniform deceleration formula based on the maximum deceleration output by the AGV driver and the speed before braking. The minimum safe braking distance is obtained by summing the braking distances of the first and second parts.

[0075] Ultrasonic sensors are used to determine the near-field area around the AGV. When an inaccessible obstacle is detected, the corresponding directional angle interval is marked as a prohibited sector and removed from the passable sectors. Specifically, ultrasonic sensors are used to detect the closest distance of obstacles in the near-field areas in front of, to the side and diagonally to the AGV. When the closest distance is less than the minimum braking distance required at the current speed, it is determined that the corresponding directional angle interval does not meet the safe approach conditions and is marked as a prohibited sector.

[0076] By conditionally merging the passable sectors with the validity level of the AGV's current reference heading, a navigation constraint field is ultimately formed that simultaneously meets spatial safety requirements and uses the BeiDou heading as a long-term stable constraint. Specifically, when the validity level of the AGV's current reference heading is valid, the AGV travels within the passable sector according to the reference heading. This enhances the vehicle's ability to follow the BeiDou heading reference during travel while ensuring spatial safety, thereby achieving long-term heading stability. When the validity level of the AGV's current reference heading is invalid, no directional constraint on the passable sector is applied. Navigation decisions are made entirely based on the passable sector constructed by lidar and ultrasonic sensors, thus maintaining the continuity and safety of navigation behavior in the event of BeiDou navigation failure.

[0077] It should be noted that by conditionally fusing the passable sectors constructed by lidar and ultrasonic sensors with the effectiveness level of the AGV's current reference heading, a navigation constraint field is formed with the BeiDou heading as a long-term stable constraint and local environmental safety as an immediate limitation. This allows the AGV to maintain heading stability and adapt to changes in BeiDou heading quality without relying on positioning and mapping, thereby improving the stability and continuity of the navigation process.

[0078] S4. Based on the navigation constraint field, before the AGV enters the preset shared passage area, calculate the entry time window according to the current speed of the AGV, and reserve the entry time window under unified time synchronization to generate the shared passage area driving instruction.

[0079] The shared passage area includes turnaround points, narrow passages, and intersections.

[0080] Before entering the shared passage area, based on the heading range limited to the AGV in the navigation constraint field and the AGV's current speed, the entry time window for the AGV to reach the entrance of the shared passage area from its current position along the allowed heading is estimated. This includes the earliest arrival time that the AGV can reach under normal and safe travel conditions, and the latest entry time allowed without affecting safe travel.

[0081] The entry time window is compared with the time window occupancy table corresponding to the shared passage area. The time window occupancy table records the entry time windows of the shared passage area that have been occupied or reserved by other AGVs under the unified time axis of Beidou time synchronization. The entry time windows that do not overlap with the entry time windows of other AGVs are filtered out by the determination method of non-overlapping time intervals.

[0082] During the entry time window screening process, the start-up delay time and braking response time that are objectively present in the actual operation of AGV are superimposed on the entry time window as fixed safety time conditions to ensure that AGV has the necessary physical reaction time when entering or leaving the shared passage area, thereby avoiding the time window non-executability problem caused by the assumption of ideal instantaneous entry or exit.

[0083] When there are entry time windows in the time window occupancy table that meet the filtering conditions, the earliest entry time window is reserved, and a shared passage area driving instruction is generated with a fixed start time for entering the shared passage area and a mandatory exit time.

[0084] When there is no entry time window in the time window occupancy table that meets the filtering conditions, the system proceeds to a safe waiting position while keeping the navigation constraint field output constraints unchanged. At the same time, it releases the occupancy request for the shared passage area and re-executes the entry time window filtering process to apply for the next earliest entry time window.

[0085] It should be noted that by implementing lightweight entry time window occupancy constraints on the shared passage area, the risk of local conflicts among multiple AGVs is effectively suppressed without changing the main navigation structure. This ensures the safety and continuity of the navigation process while avoiding the increased system complexity caused by over-reliance on global scheduling in existing technologies.

[0086] S5. Based on the navigation constraint field and the driving instructions of the shared passage area, the three-dimensional instruction skeleton is corrected in course and superimposed with time window constraints under the unified time axis, and continuously expanded and updated to generate AGV navigation instruction flow.

[0087] The ternary command skeleton is corrected using a navigation constraint field. Specifically, the next target direction angle in the ternary command skeleton is mapped to the heading range covered by the passable sector. When the next target direction angle in the ternary command skeleton falls into the heading range covered by the passable sector, the next target direction angle in the ternary command skeleton remains unchanged. When the next target direction angle in the ternary command skeleton does not fall into the heading range covered by the passable sector, the heading direction with the smallest deviation from the next target direction angle in the ternary command skeleton is selected from the passable sector, and the selected heading direction replaces the next target direction angle in the ternary command skeleton.

[0088] Meanwhile, when the validity level of the AGV's current reference heading is valid, the AGV will travel based on the current reference heading and will be subject to the constraints of the passable sectors, that is, it will travel around or avoid obstacles. When the validity level of the AGV's current reference heading is invalid, the AGV will only be subject to the constraints of the passable sectors and will not be subject to the constraints of the reference heading.

[0089] By inserting shared passage area driving instructions into the ternary instruction skeleton, timing boundary constraints are applied to the ternary instruction skeleton. Specifically, when the shared passage area driving instruction is empty, the ternary instruction skeleton only retains the expected time requirements of each target station contained in the instruction skeleton itself; when the shared passage area driving instruction is not empty, the entry time and exit deadline time in the shared passage area driving instruction are superimposed on the ternary instruction skeleton, so that the ternary instruction skeleton simultaneously contains the entry time constraint and the exit deadline time constraint of the shared passage area.

[0090] After completing the correction of the navigation constraint field and inserting the driving instructions for the shared passage area, the three-dimensional instruction skeleton is combined with the next target direction angle, expected time table, and entry and exit times of the shared passage area at each consecutive moment under unified timing to form a navigation instruction, and is continuously output in the order of the unified timing time axis to obtain the navigation instruction stream of the AGV.

[0091] During the navigation command stream output process, the updates of the navigation constraint field and the shared access area driving command continue to serve as timing consistency constraints. Specifically, when the output of the navigation constraint field changes, the latest accessible sector of the navigation constraint field is used to re-execute the next target azimuth mapping and replacement process. When the change in the output of the navigation constraint field causes the entry time and exit deadline of the shared access area driving command to be unable to be met, the process returns to the shared access area driving command application process to obtain a new shared access area driving command.

[0092] It should be noted that by continuously integrating the instruction skeleton, navigation constraint field output, and shared regional time window instructions into a real-time updated navigation instruction stream under the unified time axis of BeiDou timing, the AGV can achieve safe navigation with stable heading, controllable timing, and adaptive capabilities without relying on global replanning.

[0093] This embodiment also provides a computer device applicable to the AGV adaptive navigation method based on multi-sensor fusion, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the AGV adaptive navigation method based on multi-sensor fusion as proposed in the above embodiment.

[0094] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0095] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the AGV adaptive navigation method based on multi-sensor fusion as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0096] In summary, this invention achieves the following: First, it utilizes BeiDou second pulses to establish unified timing, aligning perception, control, and multi-vehicle time window occupancy tables on the same time axis. This reduces the risk of timing drift during multi-source data fusion and command issuance. Second, it uses dual antennas to calculate an absolute reference heading consistent with true north. Third, it conditionally fuses the passable sectors constructed by lidar and ultrasound with the heading effectiveness level. This enhances long-term heading stability when the heading is effective, while degenerating into purely local safety decisions when the heading is ineffective, ensuring continuity and safety. Fourth, it reserves entry time windows based on unified timing and generates window commands before entering shared passage areas, lightweightly coupling local passage with multi-vehicle timing mutual exclusion, reducing conflicts without introducing heavy global scheduling complexity. Finally, it continuously expands and updates the ternary command skeleton to form a real-time updatable navigation command stream, achieving a unified approach to AGV heading stability, timing control, and adaptive navigation.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An AGV adaptive navigation method based on multi-sensor fusion, characterized in that: include, The Beidou receiver and dual-antenna structure are used to calculate the unified timing and reference heading of the AGV, and the navigation task of the AGV is converted into a three-element command skeleton. Based on the three-dimensional command skeleton, the inertial measurement unit is used to determine the validity of the reference heading and generate the validity level of the AGV's current reference heading; By using lidar and ultrasonic sensors to determine the nearest obstacle distance in the AGV's forward direction area, a passable sector is constructed, and this is fused with the effectiveness level of the AGV's current reference heading to obtain a navigation constraint field. Based on the navigation constraint field, before the AGV enters the preset shared passage area, the entry time window is calculated according to the current speed of the AGV, and the entry time window is reserved under unified timing to generate the driving instruction for the shared passage area; Based on the navigation constraint field and the driving instructions of the shared passage area, the ternary instruction skeleton is subject to heading correction and time window constraint superposition under a unified time axis, and is continuously expanded and updated to generate AGV navigation instruction flow.

2. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for converting the AGV's navigation task into a three-element instruction skeleton are as follows. The navigation task of the AGV is broken down into a sequence of target stations and a desired timetable; Calculate the next target direction angle based on the spatial coordinate difference between the AGV's current position and the next target station in the target station sequence; The next target direction angle, the expected timetable, and the AGV's reference heading are combined and encapsulated according to the target station sequence to obtain the ternary instruction skeleton.

3. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for determining the validity of the reference heading are as follows. The navigation control cycle is set during the AGV's travel between adjacent target stations; The angular velocity of the AGV's geometric center is measured by an inertial measurement unit, and the angular velocity is integrated within the navigation control cycle to obtain the AGV's heading change. At the same time, the AGV's reference heading change is calculated. During the navigation control cycle, the difference between the change in heading and the change in the AGV's reference heading is calculated to obtain the heading difference. A heading difference threshold is set based on the deviation between the heading change and the AGV reference heading. If the heading difference exceeds the heading difference threshold in multiple consecutive navigation control cycles, the validity level of the AGV's current reference heading is determined to be invalid. Conversely, if the heading difference does not exceed the heading difference threshold in multiple consecutive navigation control cycles, the validity level of the AGV's current reference heading is determined to be valid.

4. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for constructing a passable sector are as follows. Based on the spatial point cloud information output by the lidar, calculate the nearest distance to obstacles in the direction the AGV is moving, and divide the passable sectors of the AGV. Ultrasonic sensors are used to detect obstacles in the nearby area around the AGV, and areas with obstacles are marked as prohibited sectors. Prohibited sectors are then removed from the passable sectors.

5. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The validity level of the passable sector is integrated with the AGV's current reference heading. The specific steps are as follows: When the validity level of the current reference heading of the AGV is valid, the AGV travels within the passable sector according to the reference heading; When the validity level of the current reference heading of the AGV is invalid, no directional constraint on the reference heading is applied to the passable sector.

6. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for reserving an entry time window under unified time synchronization are as follows: The entry time window is compared with the time window occupancy table of the shared passage area. By judging the non-overlapping time intervals, the entry time windows that do not overlap with the entry time windows of other AGVs are selected. During the selection process for entering the time window, the AGV's start-up delay time and braking response time are added to the entry time window as fixed safety time conditions. When there are entry time windows in the time window occupancy table that meet the filtering criteria, the entry time window with the earliest start time of the appointment is selected. When there is no entry time window in the time window occupancy table that meets the filtering criteria, the AGV moves to a safe position to wait, and at the same time, applies for the next earliest entry time window.

7. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for applying heading correction and time window constraint superposition to the ternary command skeleton are as follows. The ternary command skeleton is modified using a navigation constraint field; The shared passage area driving instruction is inserted into the ternary instruction skeleton, and timing boundary constraints are applied to the ternary instruction skeleton.

8. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 7, characterized in that: The specific steps for modifying the ternary command skeleton using a navigation constraint field are as follows. Map the next target heading angle in the ternary command skeleton to the heading range covered by the traversable sector; When the next target direction angle in the ternary command skeleton enters the heading range covered by the passable sector, the next target direction angle in the ternary command skeleton remains unchanged. When the next target heading angle in the ternary command skeleton does not fall within the heading range covered by the passable sector, select the heading direction with the smallest deviation from the next target heading angle in the ternary command skeleton from the passable sector, and replace the next target heading angle in the ternary command skeleton with the selected heading direction.

9. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 7, characterized in that: The specific steps for applying timing boundary constraints to the ternary instruction skeleton are as follows: When the shared passage area driving instruction is empty, the three-dimensional instruction skeleton only retains the requirements of the expected timetable; When the shared passage area driving instruction is not empty, the entry time and exit deadline of the shared passage area driving instruction are superimposed on the three-element instruction skeleton.

10. The AGV adaptive navigation method based on multi-sensor fusion as described in claim 1, characterized in that: The specific steps for generating the AGV navigation command stream are as follows: After completing the course correction and time window constraint superposition, the next target direction angle, expected time table, and entry and exit times of the shared passage area at each moment are combined into a navigation command under unified timing, and continuously output in the order of the unified timing time axis to obtain the navigation command flow of the AGV.

Images