Intelligent scheduling system and method for multi-agv cooperative carrying

Abstract

The application discloses a kind of multi-AGV collaborative handling intelligent scheduling system and method, it is related to automatic guided vehicle collaborative scheduling technical field, including obtaining target object bottom support space boundary, entry opening, support position, final support posture and AGV's car body contour and motion constraint information, constructs initial support space state;Generate terminal access channel corresponding to support position, establish preposition relationship according to the influence of support occupancy envelope on the reachable state of the remaining channel, form support entry dependent chain and control AGV entry;In the process of entry, according to the actual occupancy envelope of current AGV, identify channel blocking, determine channel release posture along actual driving track, and update support entry dependent chain after channel recovery.It solves the technical problems that the existing multi-AGV collaborative entry is difficult to determine reasonable order, and lacks recovery and order adjustment means when subsequent channel is blocked due to occupancy deviation, improves the continuity and reliability of collaborative entry.

Description

Technical Field

[0001] This invention relates to the field of automated guided vehicle (AGV) collaborative scheduling technology, specifically to an intelligent scheduling system and method for multi-AGV collaborative handling. Background Technology

[0002] With the development of Automated Guided Vehicle (AGV) technology, collaborative handling by multiple AGVs has been gradually applied to the transfer of large equipment, carrying platforms, and large components. In some collaborative handling scenarios, multiple AGVs need to enter the support space through the entry opening at the bottom of the target object and move to different pre-set support positions to jointly support the target object. Since the support space at the bottom of the target object is usually limited, and the final support posture, AGV body outline, and motion constraints corresponding to each support position are different, the spatial areas that each AGV needs to pass through when approaching the corresponding support position may intersect or partially overlap. Existing scheduling methods usually control each AGV to enter the position sequentially based on the task allocation order, the distance between the AGV and the support position, or a pre-set priority, or plan the travel path from the entry position to the corresponding support position for each AGV.

[0003] However, existing scheduling methods typically focus on determining whether a single AGV can reach its corresponding support position. They struggle to accurately analyze the impact of a single AGV's occupied area on the passage conditions of other AGVs approaching the support position when multiple AGVs sequentially enter the support space at the bottom of the target object. This can easily lead to situations where the first AGV to enter occupies the passage area required by subsequent AGVs, preventing them from reaching the corresponding support position. Furthermore, the actual parking position and vehicle orientation of the AGVs may deviate from the predetermined final support posture due to positioning or motion control errors. The actual occupancy of multiple AGVs can also have a cumulative effect, causing a pre-planned passable path to become impassable during actual execution. Existing technologies lack effective means to determine a reasonable entry sequence based on the spatial occupancy impact between support positions, and to identify, restore, and adjust the remaining entry sequence when actual occupancy obstructs the passageway, thus affecting the continuity and reliability of multi-AGV collaborative entry.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide an intelligent scheduling system and method for multi-AGV collaborative handling. This aims to solve the technical problems in existing multi-AGV collaborative placement processes, such as the difficulty in determining a reasonable placement sequence based on the impact of each AGV's completion of its support and positioning on the approach channels of other AGVs, and the lack of effective means for channel identification, restoration, and adjustment of the remaining placement sequence when the approach channels of subsequent terminals are obstructed due to deviations in the actual parking position or vehicle direction of the AGVs. These issues affect the continuity and reliability of multi-AGV collaborative placement.

[0006] The objective of this invention can be achieved through the following technical solutions: An intelligent scheduling method for collaborative transport of multiple AGVs includes: S1, acquiring spatial boundary information of the bottom support space of the target object, an entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transport and corresponding one-to-one with each support position, vehicle body contour information and motion constraint information of each AGV, and a preset safety interval, and constructing an initial support space state including a passable area based on the spatial boundary information, the entry opening and the multiple support positions; S2. For each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position, generate a terminal approach channel that connects the entry opening to the corresponding final support posture. S3. For any of the support positions, generate a support occupancy envelope based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture. Superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state. Determine the reachability of the terminal approach channels corresponding to the other support positions before and after the superposition of the support occupancy envelope. When the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from a reachable state to an inaccessible state, establish an entry-precedence relationship between the second support position and the first support position. S4. Generate a support entry dependency chain based on the entry prerequisite relationship between each support position, and control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support entry dependency chain.

[0007] As a further aspect of the present invention: in step S2, the generation of the terminal proximity channel includes: Using the final support posture of the corresponding support position as the starting state for reverse analysis, multiple candidate posture nodes are generated towards the entry opening based on the vehicle body contour information and motion constraint information of the corresponding AGV. Based on the connection relationship between adjacent candidate posture nodes verified by motion constraints, a feasible posture sequence connecting the entry opening and the final support posture is generated. Based on the vehicle body sweeping area formed by the corresponding AGV moving along the feasible posture sequence, the vehicle body sweeping area is expanded outward according to the preset safety interval to obtain the terminal approach channel of the corresponding support position.

[0008] As a further aspect of the present invention: in step S3, the generation of the support occupancy envelope includes: Based on the final support posture of the corresponding support position and the vehicle body contour information of the corresponding AGV, determine the vehicle body projection area of ​​the AGV when it is in the final support posture; The vehicle body projection area is expanded outward according to the preset safety interval to obtain the support occupancy envelope of the corresponding support position.

[0009] As a further aspect of the present invention: in step S3, the reachability state of the terminal approaching the channel is determined in the following way: The impassable area corresponding to the support occupancy envelope is deducted from the terminal approach channel to be judged to form the remaining channel area after occupancy; Search within the remaining channel area after the occupancy of the space; retrieve a continuous posture sequence that extends from the entry opening to the corresponding final support posture and satisfies the corresponding AGV motion constraint information. When the continuous attitude sequence is retrieved, it is determined that the corresponding terminal approach channel is in an reachable state; when the continuous attitude sequence is not retrieved, it is determined that the corresponding terminal approach channel is in an unreachable state.

[0010] As a further aspect of the present invention: in step S4, the generation of the supporting in-position dependency chain includes: Each of the aforementioned support positions is determined as a support relationship node. Based on each of the aforementioned preceding support positions, a directed connection is established from the support relationship node corresponding to the preceding support position to the support relationship node corresponding to the following support position, forming a directed support entry relationship. The preceding support position corresponding to each support position is then determined. When there is no closed dependency path in the support-in directed relationship, the support relationship nodes are topologically layered according to the incoming connection state of each support relationship node, and the support-in dependency chain is generated according to the formation order of each topological layer.

[0011] As a further aspect of the present invention, the method further includes: S5. During the placement process according to the support placement dependency chain, the AGV that has reached the corresponding final support posture is determined as the current AGV, and the actual occupancy envelope of the current AGV is generated according to the actual parking position, actual vehicle direction and vehicle outline information of the current AGV and the preset safety interval. Based on the actual occupancy envelope of each AGV that completed its entry and remained in the support space before the current AGV occupies its position, the current support space state before the current AGV occupies its position is formed. The support occupancy envelope of the current AGV and the actual occupancy envelope of the current AGV are superimposed on the current support space state to form the pre-calculated occupancy support space state and the actual occupancy support space state. The pre-calculated reachability state and the actual reachability state of the terminal approach channel corresponding to each support position that has not yet been placed are determined respectively. When the actual reachability of each terminal approaching the channel is reachable, it is confirmed that the current AGV has completed the support position. If there are still support positions that have not been completed, the next support position to be entered is determined from the support positions that have been completed and occupied by the corresponding previous support positions. When at least one of the terminals approaches a channel whose actual reachability is unreachable, a channel blocking result is generated, channel recovery processing is performed on the current AGV, and the unexecuted part of the support placement dependency chain is updated according to the support space status after channel recovery.

[0012] As a further aspect of the present invention: the channel blocking result is generated based on the pre-calculated reachable state and the actual reachable state, wherein: The terminal approach channel that is pre-calculated as reachable but is actually unreachable is identified as the actual occupancy deviation blocking channel. Terminal approach channels that are unreachable in both the pre-calculated reachable state and the actual reachable state are identified as cumulative occupancy blocking channels. The channel blocking results include the blocked terminal approaching the channel and its corresponding blocking type.

[0013] As a further aspect of the present invention: the channel recovery process includes: The actual driving trajectory formed when the current AGV enters the support space is obtained, and multiple candidate retreat postures are determined along the actual driving trajectory; Based on the vehicle body contour information of the current AGV in each candidate retraction posture, a retraction occupancy envelope is generated, and the actual occupancy envelope in the actual occupancy support space state is replaced with each of the retraction occupancy envelopes to obtain multiple retraction simulated support space states. From the candidate retreat postures that restore the access channel of each blocked terminal to a reachable state, the candidate retreat posture with the smallest path distance along the actual driving trajectory from the actual parking position is selected as the channel release posture, and the current AGV is controlled to move to the channel release posture.

[0014] As a further aspect of the present invention: the update of the unexecuted portion of the support input dependency chain includes: The support position corresponding to the current AGV is redefined as the support position that has not yet been placed, and the actual occupancy envelope in the actual occupancy support space state is replaced with the retreat occupancy envelope when the current AGV is in the channel release posture, to form the support space state after the channel is restored. For each support position that has not yet been placed, the support occupancy envelope corresponding to the support position is simulated in the support space state after the channel is restored, and the influence of the support occupancy envelope on the terminal approaching the channel reachable state of the other support positions that have not yet been placed is determined. When simulating the support position corresponding to the current AGV, the retreat occupancy envelope is first removed from the support space state after the channel is restored. Based on the change in the reachable state, the pre-positioning relationship between each support position that has not yet been positioned, the corresponding pre-positioning support position and the support positioning dependency chain are re-determined, and the corresponding AGV is controlled to continue to perform the positioning according to the regenerated support positioning dependency chain.

[0015] Furthermore, to achieve the above objectives, the present invention also provides an intelligent scheduling system for multi-AGV collaborative handling, comprising: The support space state construction module is used to acquire the spatial boundary information of the support space at the bottom of the target object, the entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transportation and corresponding to each support position, the vehicle body contour information and motion constraint information of each AGV, and the preset safety interval, and construct an initial support space state including a passable area based on the spatial boundary information, the entry opening and multiple support positions. The terminal approach channel generation module is used to generate a terminal approach channel that connects the entry opening to the corresponding final support posture for each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position. The occupancy relationship processing module is used to generate a support occupancy envelope for any of the support positions based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture, superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state, and determine the reachability of the terminal approach channels corresponding to the other support positions before and after the support occupancy envelope is superimposed; it is also used to establish an entry-precedence relationship between the second support position and the first support position when the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from a reachable state to an inaccessible state. The in-position dependency processing module is used to generate a support in-position dependency chain based on the in-position precondition relationship between each of the support positions. The positioning control module is used to control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support positioning dependency chain.

[0016] The beneficial effects of this invention are: (1) This invention generates terminal approach channels corresponding to each support position based on the spatial boundary of the support space at the bottom of the target object, the entry opening, multiple support positions, the final support posture corresponding to each support position, and the vehicle body contour information and motion constraint information of each AGV. By simulating the influence of the support occupancy envelope formed by each AGV at the corresponding support position on the reachability of other terminal approach channels, the pre-entry relationship between each support position is established. Thus, spatial occupancy conflicts between different support positions can be identified before the AGV actually enters the support space, and a support entry dependency chain can be generated accordingly. This avoids the situation where the AGV that enters first blocks the subsequent AGV from approaching the corresponding support position due to an unreasonable entry order, thereby improving the orderliness and executability of the multi-AGV collaborative entry process. (2) In the process of AGV performing placement according to the support placement dependency chain, the present invention generates an actual occupancy envelope based on the current actual parking position, actual vehicle direction, and vehicle outline information of the AGV, and superimposes it with the support occupancy envelope of the corresponding support position onto the current support space state, thereby obtaining the pre-calculated reachable state and the actual reachable state. By comparing the two reachable states, it is possible to distinguish between actual occupancy deviation blocking and cumulative occupancy blocking, avoiding the neglect of actual parking deviation and the cumulative occupancy of multiple vehicles when judging the channel reachability based solely on the predetermined final support posture, thus improving the accuracy of channel blocking identification; (3) When it is detected that at least one terminal approach channel corresponding to an unreachable support position is in an unreachable state, the present invention can determine candidate retreat postures along the actual travel trajectory of the current AGV, select the channel release posture that makes each blocked terminal approach channel reachable again and has a smaller retreat distance, and re-determine the entry prerequisite relationship and support entry dependency chain between the remaining support positions after the channel is restored. Thus, the blocked channel can be released and subsequent entry can continue without terminating the entire collaborative entry process, improving the continuity, adaptability and reliability of the multi-AGV collaborative handling process. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method framework structure of the present invention; Figure 2 This is a schematic diagram of the system framework structure of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Please refer to Figure 1 This application provides an intelligent scheduling method for collaborative handling of multiple AGVs, including: S1. Obtain the spatial boundary information of the bottom support space of the target object, the entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transportation and corresponding to each support position, the vehicle outline information and motion constraint information of each AGV, and the preset safety interval, and construct an initial support space state including a passable area based on the spatial boundary information, the entry opening and multiple support positions. In this embodiment, the target object is a device, a support platform, or a large component that requires coordinated support from the bottom by multiple AGVs. A support space is formed between the bottom of the target object and the working ground to accommodate the entry and movement of multiple AGVs.

[0020] First, obtain the spatial boundary information of the support space.

[0021] The spatial boundary information is used to represent the space area at the bottom of the target object that allows the AGV to enter and move. The spatial boundary information is determined by the structural dimensions of the target object, the arrangement of its bottom components, and the distance between the target object and the working surface. In specific implementation, a unified planar coordinate system is established on the working plane where the bottom of the target object is located, and the boundary position of the support space is represented in this planar coordinate system. The spatial boundary information is represented by multiple sequentially connected boundary points, with adjacent boundary points forming boundary line segments, and these boundary line segments collectively defining the outer perimeter of the support space.

[0022] In this embodiment, the support space at the bottom of the target object is rectangular, with a length of eight meters and a width of six meters. The spatial boundary information includes the boundary positions of the four sides of the rectangular support space. The area outside the boundary of the support space is not within the area that the AGV can enter.

[0023] Furthermore, an access opening communicating with the support space is obtained.

[0024] The entry opening indicates the location where the AGV enters the support space from outside the target object. The entry opening is located on the spatial boundary of the support space and communicates with the interior of the support space. Information about the entry opening includes its boundary location, start position, end position, and width. Based on the start and end positions, the specific range of the entry opening on the boundary of the support space is determined.

[0025] In this embodiment, the entry opening is located at the center of one side boundary of the support space, and the opening width is 2.2 meters. The boundary portions on both sides of the entry opening still belong to the support space boundary, and the AGV can only enter the support space through the entry opening.

[0026] Furthermore, multiple support positions are obtained within the support space.

[0027] The support positions represent the locations where the AGV supports the target object after it has completed its insertion. Each support position is set at the bottom of the target object where support is required, and each has a corresponding support position identifier and spatial location. In this embodiment, four support positions are set within the support space, namely the first support position, the second support position, the third support position, and the fourth support position. The four support positions are spaced apart along the lateral direction of the bottom of the target object.

[0028] The first support position is located on the left side of the bottom of the target object, the second support position is located to the right of the first support position, the third support position is located to the right of the second support position, and the fourth support position is located on the right side of the bottom of the target object. Each support position is located within the support space defined by the spatial boundary and has a corresponding spatial distance from the entry opening.

[0029] Furthermore, the final support posture corresponding to each support position is obtained.

[0030] The final support posture refers to the position and vehicle orientation of the corresponding AGV when it reaches the support position and forms support. Each final support posture corresponds to a support position. The final support posture includes the position of the corresponding AGV's vehicle body reference point and the longitudinal direction of the vehicle body. The vehicle body reference point is the geometric center of the AGV's vehicle body.

[0031] In this embodiment, the first support position corresponds to the first final support posture, the second support position corresponds to the second final support posture, the third support position corresponds to the third final support posture, and the fourth support position corresponds to the fourth final support posture.

[0032] In all four final support postures, the vehicle body is oriented towards the same side of the bottom of the target object. The position of each final support posture is located at the center of its corresponding support position.

[0033] For example, the final support posture corresponding to the first support position is located at a distance of 1.8 meters from the left boundary of the support space and 4.6 meters from the boundary of the entrance opening; the final support postures corresponding to the second, third and fourth support positions are set sequentially at horizontal intervals.

[0034] Furthermore, multiple AGVs corresponding to each support position are obtained.

[0035] In this embodiment, the AGVs participating in the collaborative transport include a first AGV, a second AGV, a third AGV, and a fourth AGV. The first AGV corresponds to the first support position, the second AGV corresponds to the second support position, the third AGV corresponds to the third support position, and the fourth AGV corresponds to the fourth support position. Each AGV corresponds to only one support position in this collaborative transport task and corresponds to the final support posture of that support position. This establishes a one-to-one correspondence between support positions, final support postures, and AGVs.

[0036] Furthermore, obtain the vehicle body outline information of each AGV.

[0037] The vehicle body contour information is used to represent the shape of the AGV's body on the working plane and the plane range occupied by the body. The vehicle body contour information is determined based on the AGV's body length, body width, and the position of the body reference point.

[0038] When the projection of the AGV's body onto the working plane is rectangular, the body contour information includes the body length, body width, and the distance from the body reference point to each edge of the body. When the AGV's body projection is not a regular rectangle, the body contour information is represented by connecting multiple contour points sequentially.

[0039] In this embodiment, the vehicle body projections of the first AGV, second AGV, third AGV, and fourth AGV are all rectangular. The vehicle body length of each AGV is 1.2 meters, the vehicle body width is 0.8 meters, and the vehicle body reference point is located at the center of the vehicle body projection area.

[0040] After determining the specific posture of the AGV within the support space, the actual vehicle projection area of ​​the AGV within the support space is determined based on the position of the vehicle reference point, the vehicle direction, and the vehicle outline information.

[0041] Furthermore, the motion constraint information of each AGV is obtained.

[0042] The motion constraint information is used to represent the motion modes that the AGV can perform within the support space. The motion constraint information is determined based on the AGV's own driving structure and steering capability. The motion constraint information includes the AGV's allowed driving direction, allowed steering mode, minimum turning radius, and the allowed range of directional changes between adjacent motion postures.

[0043] In this embodiment, each AGV is capable of both forward and reverse movement, and can turn during operation. The minimum turning radius of each AGV is 0.7 meters.

[0044] When an AGV continuously traverses multiple motion postures, the change in vehicle direction between adjacent postures does not exceed ten degrees. Therefore, as the AGV's position changes, its vehicle direction changes continuously without abrupt shifts in position or direction. The motion constraint information of each AGV is associated with its corresponding AGV. The motion constraint information of the first AGV corresponds to the first AGV, the motion constraint information of the second AGV corresponds to the second AGV, and the third and fourth AGVs use the same association method.

[0045] Furthermore, obtain the preset safety interval.

[0046] The preset safety interval represents the safe distance reserved outside the AGV's body outline. This preset safety interval is used to compensate for AGV positioning errors, motion errors, and the necessary safe distance between the vehicle body and surrounding structures. The preset safety interval is pre-set based on the AGV's positioning accuracy, body size, and actual passage conditions of the support space.

[0047] In this embodiment, the preset safety interval is 0.15 meters. This preset safety interval applies to all AGVs participating in this collaborative transport.

[0048] After acquiring the above information, the initial support space state is constructed based on the spatial boundary information, entry opening, multiple support positions, the final support posture corresponding to each support position, multiple AGVs corresponding to each support position, the vehicle body contour information and motion constraint information of each AGV, and the preset safety interval.

[0049] Specifically, the overall range of the support space is determined based on the spatial boundary information, and the area outside the spatial boundary is defined as a non-passable area. The area inside the spatial boundary and connected to the entrance opening is defined as a passable area. The passable area represents the area within the support space that can be used for AGV motion analysis when no AGV occupancy area is superimposed.

[0050] The entry opening, serving as a connection point between the support space and the external area, is included in the initial support space state. Each support position and its corresponding final support posture, as the target position and target posture within the support space, are also included in the initial support space state. The one-to-one correspondence between each support position and its corresponding AGV, the vehicle body contour information of each AGV, the motion constraint information of each AGV, and the preset safety interval are associated with the corresponding support position and the final support posture, respectively. In the initial support space state, each support position is not yet occupied by its corresponding AGV. Therefore, the initial support space state does not include the support-occupying area formed by the AGV at the support position, nor does it include the actual occupying area formed after the AGV actually stops. This forms the initial support space state, which reflects the initial spatial situation of the target object's bottom support space before each AGV enters and occupies its corresponding support position.

[0051] S2. For each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position, generate a terminal approach channel that connects the entry opening to the corresponding final support posture. Specifically, in this embodiment, a first support position, a second support position, a third support position, and a fourth support position are provided within the support space, each support position corresponding to a first AGV, a second AGV, a third AGV, and a fourth AGV, respectively. The same terminal approach channel generation process is performed for each support position, and the terminal approach channels corresponding to each support position are determined independently.

[0052] For the first support position, the final support posture corresponding to the first support position is determined as the starting state for reverse analysis. The final support posture includes the vehicle body reference point position and vehicle body orientation when the first AGV completes its insertion. Taking the geometric center of the first AGV's vehicle body as the vehicle body reference point, the position and vehicle body orientation of the first AGV in the final support posture are defined as a specific posture node.

[0053] Starting from the final support posture, multiple candidate posture nodes are generated progressively in the direction of the entry opening. Each candidate posture node includes the position of the corresponding AGV's vehicle body reference point and the vehicle body orientation. A candidate posture node represents a possible vehicle body state that the corresponding AGV may traverse during its journey from the entry opening to the final support posture. When generating candidate posture nodes, the permissible positional and directional changes between adjacent candidate posture nodes are determined based on the motion constraint information of the first AGV.

[0054] In this embodiment, the first AGV can travel forward and backward, can turn during travel, has a minimum turning radius of 0.7 meters, and the change in vehicle direction between adjacent candidate posture nodes does not exceed 10 degrees.

[0055] Using the final support posture as the current posture node, generate the previous candidate posture node along the direction closest to the entry opening. In one scenario of generating a candidate posture node, keep the vehicle orientation unchanged and move the vehicle reference point longitudinally along the vehicle body; in another scenario, under the condition of satisfying the minimum turning radius, move the vehicle reference point along the arc direction and change the vehicle orientation accordingly.

[0056] In one specific implementation, the positional interval between adjacent candidate posture nodes is set to 0.2 meters. For each candidate posture node generated, the vehicle projection area of ​​the first AGV when it is located at the candidate posture node is determined based on the position of the vehicle reference point and the vehicle orientation in the candidate posture node, combined with the vehicle contour information of the first AGV.

[0057] The first AGV has a rectangular body outline, with a length of 1.2 meters and a width of 0.8 meters. The center position of the vehicle projection area is determined based on the vehicle reference point position in the candidate posture node, and the orientation of the vehicle projection area is determined based on the vehicle orientation in the candidate posture node, thus obtaining the planar area occupied by the first AGV under the corresponding candidate posture node.

[0058] The vehicle projection area is compared with the passable area in the initial support space. When the entire vehicle projection area of ​​the first AGV at a candidate posture node is within the passable area, the candidate posture node is determined to meet the spatial boundary requirements. When the vehicle projection area of ​​the first AGV at a candidate posture node exceeds the passable area or overlaps with the impassable area in the initial support space, the candidate posture node is discarded. For candidate posture nodes that meet the spatial boundary requirements, adjacent candidate posture nodes are generated towards the entrance opening. This process is gradually expanded to form multiple candidate posture nodes distributed from the final support posture towards the entrance opening. During the generation of candidate posture nodes, the same candidate posture node corresponds to multiple previous candidate posture nodes in different directions. For example, when the vehicle direction remains unchanged, a straight-line extending candidate posture node is formed; when the vehicle direction gradually changes, a left-turn or right-turn candidate posture node is formed.

[0059] Each candidate posture node must satisfy the condition that the projection area of ​​the corresponding AGV's vehicle body is located within the passable area. Therefore, the distribution range of candidate posture nodes is limited by the spatial boundaries and passable areas in the initial support space state. As a generated candidate posture node gradually approaches the entrance opening, it is determined whether the projection area of ​​the corresponding AGV under that candidate posture node can pass through the entrance opening. When the projection area of ​​the corresponding AGV's vehicle body is located within the passable area defined by the entrance opening, and the candidate posture node connects with the AGV's entry direction outside the entrance opening, the candidate posture node is determined as a candidate posture node connected to the entrance opening.

[0060] In this embodiment, the width of the entry opening is 2.2 meters, and the width of the AGV body is 0.8 meters. Therefore, the AGV can pass through the entry opening with its body facing the interior of the support space in the longitudinal direction.

[0061] After generating multiple candidate posture nodes, motion constraint verification is performed between adjacent candidate posture nodes. This motion constraint verification determines whether the corresponding AGV can continuously move from one candidate posture node to an adjacent candidate posture node. For two adjacent candidate posture nodes with the same vehicle orientation, it is determined whether the direction of position change between the two candidate posture nodes belongs to the allowed travel direction of the corresponding AGV, and whether the movement area between the two candidate posture nodes is located within the passable area. When the vehicle orientation between adjacent candidate posture nodes changes, the AGV's turning process is determined based on the position and direction changes between the two candidate posture nodes, and it is determined whether this turning process meets the minimum turning radius of the corresponding AGV.

[0062] In this embodiment, when the turning radius between adjacent candidate attitude nodes is less than 0.7 meters, it is determined that the two candidate attitude nodes have not passed the motion constraint verification, and no connection relationship is established between them.

[0063] When the turning radius between adjacent candidate posture nodes is not less than 0.7 meters and the change in vehicle direction does not exceed 10 degrees, it is further determined whether the AGV is always within the passable area during the process of moving from the previous candidate posture node to the next candidate posture node.

[0064] Specifically, several intermediate postures are determined between two adjacent candidate posture nodes. Each intermediate posture is located between the positions of the two candidate posture nodes, and the vehicle orientation of each intermediate posture continuously changes according to the orientation change between the two candidate posture nodes. The vehicle projection area of ​​the AGV in each intermediate posture is determined. When the vehicle projection area corresponding to each intermediate posture is within the passable area, it is determined that the AGV can move continuously between the two candidate posture nodes. When the vehicle projection area corresponding to any intermediate posture exceeds the passable area, it is determined that the AGV cannot move continuously between the two candidate posture nodes. For adjacent candidate posture nodes that pass the motion constraint verification, a connection relationship is established between them. For adjacent candidate posture nodes that fail the motion constraint verification, no connection relationship is established.

[0065] Following the above method, motion constraint verification is performed on each adjacent candidate attitude node to form a connection relationship between multiple candidate attitude nodes. This connection relationship begins with a candidate attitude node connected to the entry opening and extends sequentially according to the connection relationship between adjacent candidate attitude nodes. When the sequentially connected candidate attitude nodes can continuously extend to the final support attitude corresponding to the first support position, the group of candidate attitude nodes is arranged in the direction from the entry opening to the final support attitude, forming a feasible attitude sequence corresponding to the first support position. The first attitude node in the feasible attitude sequence is connected to the entry opening, and the last attitude node is the final support attitude corresponding to the first support position. All adjacent attitude nodes pass motion constraint verification.

[0066] The AGV body reference point position in the feasible posture sequence changes continuously, and the vehicle direction changes continuously according to the motion constraints of the corresponding AGV, enabling the first AGV to gradually move from the entry opening to the final support posture corresponding to the first support position. When there are multiple sets of connection relationships between candidate posture nodes that can continuously extend from the entry opening to the final support posture, each set of connection relationships forms a corresponding feasible posture sequence. One set of feasible posture sequences that meets the spatial boundary requirements and motion constraint information is selected as the posture sequence for generating the terminal approaching the channel corresponding to the first support position. After obtaining the feasible posture sequence corresponding to the first support position, the vehicle projection area of ​​the first AGV when it is in each posture node of the feasible posture sequence is determined sequentially. For the movement process between adjacent posture nodes, multiple intermediate postures are determined according to the continuous change of the AGV body reference point position and vehicle direction, and the vehicle projection area of ​​the first AGV when it is in each intermediate posture is determined respectively.

[0067] The entire vehicle body projection area of ​​the first AGV at each posture node and intermediate posture in the feasible posture sequence is merged to form the vehicle body sweep area when the first AGV moves along the feasible posture sequence. The vehicle body sweep area includes the planar range traversed by the vehicle body throughout the entire process of the first AGV's movement from entering the opening to its final support posture, and not just the vehicle body projection area at each discrete posture node. In the straight-line movement section, the vehicle body sweep area extends continuously along the first AGV's travel direction. In the turning movement section, the vehicle body sweep area forms corresponding arc-shaped areas according to the movement range of each edge of the first AGV's body during the turning process. After forming the vehicle body sweep area, it is expanded outward according to a preset safety interval.

[0068] In this embodiment, the preset safety interval is 0.15 meters. Based on the outer boundary of the vehicle body sweeping area, the area is extended outward by 0.15 meters, so that the extended area includes both the area actually traversed by the vehicle body and the safety area reserved outside the vehicle body.

[0069] In the section where the first AGV moves in a straight line, its body width is 0.8 meters. After adding a preset safety interval of 0.15 meters on each side of the body, the width of the corresponding area is 1.1 meters.

[0070] In the section where the first AGV turns, the width of the turning section is determined by extending outward by 0.15 meters based on the actual sweep range of the vehicle body formed when the first AGV turns. Therefore, the width of the turning section is determined by the change in vehicle body direction and the vehicle body outline.

[0071] After the outward expansion process is completed according to the preset safety interval, the resulting continuous area is determined as the terminal approach channel corresponding to the first support position. Following the same method, for the second, third, and fourth support positions, using their respective final support postures as the starting state for reverse analysis, multiple candidate posture nodes are generated towards the entry opening based on the vehicle body contour information and motion constraint information of the corresponding AGV. Motion constraint verification is performed on adjacent candidate posture nodes corresponding to each support position to form a feasible posture sequence connecting the entry opening and the corresponding final support posture. Based on the vehicle body sweep area formed when the corresponding AGV moves along its respective feasible posture sequence, outward expansion is performed according to a preset safety interval of 0.15 meters to obtain the terminal approach channels corresponding to the first, second, third, and fourth support positions. Thus, each terminal approach channel extends continuously from the entry opening to the final support posture of the corresponding support position, and each terminal approach channel matches the vehicle body contour information and motion constraint information of the corresponding AGV.

[0072] S3. For any of the support positions, generate a support occupancy envelope based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture. Superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state. Determine the reachability of the terminal approach channels corresponding to the other support positions before and after the superposition of the support occupancy envelope. When the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from a reachable state to an inaccessible state, establish an entry-precedence relationship between the second support position and the first support position. Specifically, in this embodiment, the support space is provided with a first support position, a second support position, a third support position, and a fourth support position, each corresponding to a first AGV, a second AGV, a third AGV, and a fourth AGV, respectively. For each support position, the impact of the corresponding AGV occupying that support position on the approach channel of the terminals corresponding to the other support positions is determined.

[0073] First, the first support position is used as the current simulated occupancy support position. The final support posture corresponding to the first support position and the vehicle outline information of the first AGV are obtained. The final support posture includes the vehicle reference point position and vehicle orientation when the first AGV completes its insertion. The vehicle outline information represents the shape of the first AGV's body on the working plane. Based on the final support posture, the vehicle outline of the first AGV is placed at the corresponding vehicle reference point position, and the orientation of the vehicle outline is adjusted according to the vehicle orientation in the final support posture, thereby determining the vehicle projection area of ​​the first AGV when it is in the final support posture corresponding to the first support position. The vehicle projection area represents the actual planar range covered by the first AGV's body within the support space after it completes its insertion according to the predetermined position and direction.

[0074] In this embodiment, the vehicle body projection of each AGV is rectangular, with a length of 1.2 meters and a width of 0.8 meters. When the first AGV is in the final support posture corresponding to the first support position, the length and width directions of the rectangular vehicle body projection area are determined according to the vehicle body direction of the first AGV, with the vehicle body reference point in the final support posture as the center.

[0075] After determining the vehicle body projection area, the vehicle body projection area is expanded outward according to a preset safety interval. The expansion process refers to adding an area corresponding to the preset safety interval along the outer side of the vehicle body projection area, based on the outer boundary of the vehicle body projection area.

[0076] In this embodiment, the preset safety interval is 0.15 meters. Therefore, a safety area of ​​0.15 meters is added around the projection area of ​​the first AGV to form the support occupancy envelope corresponding to the first support position.

[0077] The support occupancy envelope includes the vehicle body projection area actually occupied by the first AGV after it completes its final support posture and the safety area surrounding the vehicle body projection area.

[0078] When the direction of the AGV body is not parallel to the boundary direction of the support space, the actual orientation of the vehicle body projection area is first determined according to the final support posture, and then the area is expanded outward along the actual outer boundary of the vehicle body projection area, instead of forming a rectangular expansion area only according to the coordinate direction of the support space.

[0079] After the support occupancy envelope corresponding to the first support position is formed, the area corresponding to the support occupancy envelope is determined as an impassable area and superimposed on the initial support space state.

[0080] During the superposition process, the spatial boundaries, entry openings, original impassable areas, and passable areas in the initial support space state remain unchanged. Only the area covered by the support occupancy envelope corresponding to the first support position is adjusted from passable to impassable, thereby forming the simulated occupancy support space state corresponding to the first support position.

[0081] The simulated occupancy support space state represents the support space state when the first AGV has completed its occupancy according to the final support posture corresponding to the first support position, while the other AGVs have not yet completed their occupancy in the corresponding support positions.

[0082] When determining the impact of the support occupancy envelope corresponding to the first support position on the access channels of other terminals, the reachability states of the access channels of the terminals corresponding to the second, third, and fourth support positions before and after the support occupancy envelopes are superimposed are determined respectively.

[0083] Taking the terminal approach channel corresponding to the second support position as an example, before the support occupancy envelope corresponding to the first support position is superimposed, the terminal approach channel corresponding to the second support position extends continuously from the entry opening to the final support posture corresponding to the second support position. Therefore, the terminal approach channel is in a reachable state.

[0084] After superimposing the support occupancy envelope corresponding to the first support position, the impassable area corresponding to the support occupancy envelope is deducted from the terminal approach channel corresponding to the second support position to form the remaining channel area after occupancy corresponding to the second support position.

[0085] When the support occupancy envelope corresponding to the first support position does not overlap with the terminal approach channel corresponding to the second support position, the remaining channel area after the second support position is occupied remains consistent with the original terminal approach channel.

[0086] When the support occupancy envelope corresponding to the first support position overlaps with the terminal approach channel portion corresponding to the second support position, the overlapping portion is deducted from the terminal approach channel, and the portion not covered by the support occupancy envelope is retained as the remaining channel area after occupancy.

[0087] The fact that the support occupancy envelope partially overlaps with the terminal approach channel does not directly indicate that the terminal approach channel is in an unreachable state. It is also necessary to determine whether, after deducting the support occupancy envelope, there is still a region that allows the second AGV to move continuously from the entry opening to the final support posture corresponding to the second support position.

[0088] Within the remaining channel area after the occupancy, the continuous posture sequence is retrieved based on the vehicle body contour information and motion constraint information of the second AGV, with the entry opening as the starting area and the final support posture corresponding to the second support position as the ending state of the continuous posture sequence.

[0089] During the retrieval process, the multiple possible orientations of the second AGV are determined sequentially according to the direction from the entry opening toward the final support posture corresponding to the second support position.

[0090] Each posture includes the position of the second AGV's vehicle reference point and its orientation. Based on the second AGV's vehicle contour information, the vehicle projection area of ​​the second AGV in each posture is determined.

[0091] The vehicle body projection area corresponding to each posture should be located within the remaining channel area after the occupancy, and should not overlap with the support occupancy envelope corresponding to the first support position.

[0092] The position and orientation changes between adjacent postures should also satisfy the motion constraint information of the second AGV.

[0093] In this embodiment, the second AGV is capable of moving forward and backward, with a minimum turning radius of 0.7 meters, and the directional change between adjacent postures does not exceed ten degrees. Therefore, adjacent postures should be able to be continuously connected through linear motion or turning motion that meets the minimum turning radius.

[0094] When multiple postures that meet the above conditions can be continuously determined starting from the entry opening, and the multiple postures eventually extend to the final support posture corresponding to the second support position, it is determined that a continuous posture sequence is retrieved in the remaining channel area after the occupancy.

[0095] When the continuous attitude sequence is retrieved, it is determined that the terminal approach channel corresponding to the second support position is still in a reachable state after the support occupancy envelope corresponding to the first support position is superimposed.

[0096] If a posture that satisfies the second AGV motion constraint information cannot be continuously determined from the entry opening, or if the determined posture cannot be continuously extended to the final support posture corresponding to the second support position, it is determined that no continuous posture sequence has been retrieved.

[0097] If the continuous attitude sequence is not found, it is determined that the terminal approach channel corresponding to the second support position is in an unreachable state after the support occupancy envelope corresponding to the first support position is superimposed.

[0098] For example, the support occupancy envelope corresponding to the first support position is located in the inner passage section of the terminal approaching the channel corresponding to the second support position. When the support occupancy envelopes are superimposed, although a portion of the channel area is still retained between the entry opening and the support occupancy envelope, and a portion of the channel area is also retained between the support occupancy envelope and the final support posture corresponding to the second support position, a continuous connection that satisfies the second AGV vehicle body contour information and motion constraint information cannot be formed between the two portions of the channel area.

[0099] In this situation, the second AGV cannot move continuously from the entry opening to the final support posture corresponding to the second support position. Therefore, the terminal approach channel corresponding to the second support position changes from the reachable state before the support occupancy envelope is superimposed to the unreachable state after superposition.

[0100] When the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from an reachable state to an unreachable state, it indicates that after the first AGV completes its entry into the first support position, it will prevent the second AGV from continuing to enter the second support position.

[0101] This establishes a precedence relationship between the second support level and the first support level.

[0102] The second support position precedes the first support position, meaning that the AGV corresponding to the second support position should enter the support space and reach the final support posture corresponding to the second support position before the AGV corresponding to the first support position completes its support occupancy.

[0103] For the terminal approach channels corresponding to the third and fourth support positions, the remaining channel areas after the positions are occupied are formed in the same way, and continuous posture sequences that satisfy the motion constraint information of the third and fourth AGVs are retrieved respectively.

[0104] When a continuous attitude sequence can be retrieved, it is determined that the corresponding terminal approach channel is still in a reachable state, and no position precedence relationship is established between the corresponding support position and the first support position.

[0105] When no continuous attitude sequence is found, it is determined that the corresponding terminal approach channel changes from an reachable state to an unreachable state, and an in-position precedence relationship is established between the corresponding support position and the first support position.

[0106] After completing the simulated occupancy judgment corresponding to the first support position, the initial support space state is restored, that is, the support occupancy envelope corresponding to the first support position is no longer retained.

[0107] Subsequently, taking the second support position as the current simulated position, the vehicle projection area of ​​the second AGV when it is in the corresponding final support position is determined based on the final support posture corresponding to the second support position and the vehicle body contour information of the second AGV.

[0108] The vehicle body projection area is expanded outward according to the same preset safety interval to obtain the support occupancy envelope corresponding to the second support position.

[0109] The support occupancy envelope corresponding to the second support position is superimposed separately as an impassable area onto the initial support space state to form the simulated occupancy support space state corresponding to the second support position.

[0110] Determine the reachability of the approach channel of the terminal corresponding to the first support position, the third support position and the fourth support position before and after the support occupancy envelope is superimposed, and determine whether to establish the corresponding support position before the second support position based on whether the reachability changes from reachable to unreachable.

[0111] In the same manner, support occupancy envelope generation, simulated occupancy, and terminal approach channel reachability determination are performed for the third and fourth support positions, respectively.

[0112] In each simulated placement process, only the support placement envelope corresponding to the current support position is superimposed onto the initial support space state, without superimposing the support placement envelopes corresponding to other support positions at the same time.

[0113] Thus, the impact of each support position completing its own occupation on the approach channel of the corresponding terminals of the other support positions was obtained.

[0114] In a specific implementation, the support occupancy envelope corresponding to the second support position causes the terminal approach channel corresponding to the first support position to change from an reachable state to an unreachable state. Therefore, an in-position pre-position relationship is established where the first support position precedes the second support position.

[0115] The support occupancy envelope corresponding to the third support position causes the terminal approaching the channel corresponding to the fourth support position to change from an reachable state to an unreachable state. Therefore, an entry-preceding relationship is established where the fourth support position precedes the third support position. After the support occupancy envelopes corresponding to the remaining support positions are superimposed, a continuous attitude sequence satisfying the corresponding AGV motion constraint information can be retrieved within the remaining channel region after the corresponding occupancy. Therefore, no corresponding entry-preceding relationship is established. The resulting entry-preceding relationships include: the first support position precedes the second support position, and the fourth support position precedes the third support position.

[0116] S4. Generate a support entry dependency chain based on the entry prerequisite relationship between each support position, and control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support entry dependency chain.

[0117] Specifically, in this embodiment, the support space is provided with a first support position, a second support position, a third support position, and a fourth support position. The first support position, the second support position, the third support position, and the fourth support position correspond to the first AGV, the second AGV, the third AGV, and the fourth AGV, respectively, and each has a corresponding first terminal approach channel, a second terminal approach channel, a third terminal approach channel, and a fourth terminal approach channel.

[0118] Based on the changes in the reachability of the terminal approach channel obtained after simulating the occupancy of each support position, the pre-positioning relationship of the first support position preceding the second support position, and the pre-positioning relationship of the fourth support position preceding the third support position, have been determined.

[0119] First, each of the aforementioned support positions is designated as a support relationship node. Specifically, the first support position is designated as the first support relationship node, the second support position as the second support relationship node, the third support position as the third support relationship node, and the fourth support position as the fourth support relationship node. Each support relationship node corresponds to one support position, and with that support position, it corresponds to the corresponding AGV, the final support posture, and the terminal approach channel.

[0120] Based on the aforementioned pre-position relationships, directed connections are established between the corresponding support relationship nodes. For a pre-position relationship where the first support position precedes the second support position, a directed connection is established from the first support relationship node to the second support relationship node. The starting end of the directed connection is the first support relationship node, and the pointing end is the second support relationship node, indicating that the first support position is the pre-position of the second support position, and the first support position should be placed before the second support position. For a pre-position relationship where the fourth support position precedes the third support position, a directed connection is established from the fourth support relationship node to the third support relationship node. The starting end of the directed connection is the fourth support relationship node, and the pointing end is the third support relationship node, indicating that the fourth support position is the pre-position of the third support position, and the fourth support position should be placed before the third support position. The first support relationship node, the second support relationship node, the third support relationship node, and the fourth support relationship node, as well as the directed connections from the first support relationship node to the second support relationship node and from the fourth support relationship node to the third support relationship node, together form the support placement directed relationship.

[0121] In the directed support relationship, a directed connection pointing to a support node is the incoming connection of that support node, and a directed connection from one support node to other support nodes is the outgoing connection of that support node.

[0122] Based on the incoming connections of each supporting node, the preceding supporting position for each supporting position is determined. The first supporting node has no incoming connections; therefore, the first supporting position has no preceding supporting position. The second supporting node has an incoming connection from the first supporting node to the second supporting node; therefore, the first supporting position is the preceding supporting position for the second supporting position. The third supporting node has an incoming connection from the fourth supporting node to the third supporting node; therefore, the fourth supporting position is the preceding supporting position for the third supporting position. The fourth supporting node has no incoming connections; therefore, the fourth supporting position has no preceding supporting position.

[0123] When a support relationship node has two or more incoming connections, the support position corresponding to the starting node of each incoming connection is the incoming preceding support position of the corresponding support position of the support relationship node.

[0124] For example, when a support node has directed connections from both a first support node and a second support node, the first support bit and the second support bit are both incoming preceding support bits for the corresponding support bit of that node. The support bit needs to be located in the topology layer following the first and second support bits.

[0125] After forming the directed support relationship, it is determined whether a closed dependency path exists in the directed support relationship. The closed dependency path refers to a connection path that starts from a support node, passes through one or more other support nodes in sequence according to the direction of each directed connection, and then returns to the original support node.

[0126] For example, when there is a directed connection where the first supporting node points to the second supporting node, the second supporting node points to the third supporting node, and the third supporting node points to the first supporting node, a closed dependency path is formed between the three supporting nodes.

[0127] A closed dependency path means that each support bit in the path requires the other support bits to complete their insertion first, so the support bit that can be inserted first cannot be determined by the corresponding directed connection.

[0128] In this embodiment, the existence of a closed dependency path is determined by the incoming connection status of each supporting node.

[0129] First, search for support relationship nodes without incoming connections from all support relationship nodes. Since neither the first nor the fourth support relationship node has incoming connections, they can be extracted first. After extracting the first support relationship node, remove the directed connections from the first support relationship node to the second support relationship node. After removing these directed connections, the second support relationship node no longer has incoming connections. After extracting the fourth support relationship node, remove the directed connections from the fourth support relationship node to the third support relationship node. After removing these directed connections, the third support relationship node no longer has incoming connections. Subsequently, both the second and third support relationship nodes can be extracted. Since the first, second, third, and fourth support relationship nodes can all be extracted sequentially according to their incoming connection states, there are no closed dependency paths in the support in-position directed relationships.

[0130] After determining that there is no closed dependency path in the support-in directed relationship, topology layering is performed according to the incoming connection status of each support relationship node.

[0131] When performing the first topology layering, select the support relationship nodes that do not have incoming connections from all support relationship nodes that have not yet been divided into topology layers.

[0132] The first support node has no incoming connections, and the fourth support node also has no incoming connections. Therefore, the first support node and the fourth support node are assigned to the first topology layer.

[0133] The first topology layer includes a first support level and a fourth support level. Neither the first nor the fourth support level requires a preceding support level to be filled; therefore, the first and fourth support levels are the first support levels to satisfy the preceding filling relationship.

[0134] After the first topology layer is formed, the first support relationship node and the fourth support relationship node are removed from the current support-in directed relationship, and the directed connections issued by the first support relationship node and the fourth support relationship node are also removed.

[0135] After removing the first supporting node and the directed connections from the first supporting node to the second supporting node, the number of inbound connections to the second supporting node becomes zero.

[0136] After removing the fourth support node and the directed connections from the fourth support node to the third support node, the number of inbound connections to the third support node becomes zero.

[0137] When performing the second topology layering, neither the second support node nor the third support node has an incoming connection. Therefore, the second support node and the third support node are assigned to the second topology layer.

[0138] The second topology layer includes a second support level and a third support level.

[0139] The positional relationship between the first support position and the second support position is reflected by the fact that the first support position is located in the first topology layer and the second support position is located in the second topology layer.

[0140] The positional relationship between the fourth support position and the third support position is reflected by the fact that the fourth support position is located in the first topological layer and the third support position is located in the second topological layer.

[0141] After all supporting relationship nodes have completed topological layering, the supporting in-place dependency chain is generated in the order of first topological layer first, then second topological layer.

[0142] There are no sequential constraints between support positions located in the same topology layer due to the aforementioned pre-position relationship. When performing pre-position operations on each support position in the same topology layer, the formation order of that topology layer relative to other topology layers is not changed.

[0143] According to the support insertion dependency chain, the first AGV and the fourth AGV corresponding to the first topology layer are first controlled to enter the support space.

[0144] For the first support position, read the first terminal approach channel and the first final support posture corresponding to the first support position. Control the first AGV to enter the first terminal approach channel through the entry opening, and move along the feasible posture sequence extending from the entry opening to the first final support posture in the first terminal approach channel.

[0145] During its movement, the first AGV sequentially passes through each of the feasible postures in the sequence. The position of its vehicle reference point and the direction of its vehicle body change continuously according to the connection relationship between adjacent postures until the first AGV reaches the final support posture corresponding to the first support position, thus completing the insertion process corresponding to the first support position.

[0146] For the fourth support position, read the fourth terminal approach channel and the fourth final support posture corresponding to the fourth support position. Control the fourth AGV to enter the fourth terminal approach channel through the entry opening, and move along the corresponding feasible posture sequence in the fourth terminal approach channel until the fourth AGV reaches the final support posture corresponding to the fourth support position, completing the insertion process corresponding to the fourth support position.

[0147] After the first AGV and the fourth AGV both reach their respective final support postures, the first topology layer completes its execution.

[0148] Subsequently, the second and third support bits in the second topology layer are executed according to the support insertion dependency chain.

[0149] The preceding support position for the second support position is the first support position. Since the first AGV has reached the final support posture corresponding to the first support position, the preceding support position relationship for the second support position is satisfied.

[0150] Read the second terminal approach channel and the second final support posture corresponding to the second support position, control the second AGV to enter the second terminal approach channel from the entry opening, and move along the corresponding feasible posture sequence in the second terminal approach channel until the second AGV reaches the final support posture corresponding to the second support position.

[0151] The preceding support position for the third support position is the fourth support position. Since the fourth AGV has reached the final support posture corresponding to the fourth support position, the preceding support position relationship for the third support position is satisfied.

[0152] Read the third terminal approach channel and the third final support posture corresponding to the third support position, control the third AGV to enter the third terminal approach channel from the entry opening, and move along the corresponding feasible posture sequence in the third terminal approach channel until the third AGV reaches the final support posture corresponding to the third support position.

[0153] When a support position has multiple entry-preceding support positions, after the AGVs corresponding to each entry-preceding support position have reached their respective final support postures, the AGVs corresponding to that support position are then controlled to enter along their terminal approach channels.

[0154] Each AGV enters the support space along the corresponding terminal approach channel according to the formation order of the topology layer where its corresponding support position is located, and reaches the corresponding final support posture.

[0155] S5. During the placement process according to the support placement dependency chain, the AGV that has reached the corresponding final support posture is determined as the current AGV, and the actual occupancy envelope of the current AGV is generated according to the actual parking position, actual vehicle direction and vehicle outline information of the current AGV and the preset safety interval. Based on the actual occupancy envelope of each AGV that completed its entry and remained in the support space before the current AGV occupies its position, the current support space state before the current AGV occupies its position is formed. The support occupancy envelope of the current AGV and the actual occupancy envelope of the current AGV are superimposed on the current support space state to form the pre-calculated occupancy support space state and the actual occupancy support space state. The pre-calculated reachability state and the actual reachability state of the terminal approach channel corresponding to each support position that has not yet been placed are determined respectively. When the actual reachability of each terminal approaching the channel is reachable, it is confirmed that the current AGV has completed the support position. If there are still support positions that have not been completed, the next support position to be entered is determined from the support positions that have been completed and occupied by the corresponding previous support positions. When at least one of the terminals approaches a channel whose actual reachability is unreachable, a channel blocking result is generated, channel recovery processing is performed on the current AGV, and the unexecuted part of the support placement dependency chain is updated according to the support space status after channel recovery.

[0156] Specifically, the actual parking position of the current AGV refers to the actual position of its reference point in the support space after the current AGV stops moving; the actual vehicle direction refers to the actual orientation of the longitudinal direction of the current AGV in the support space after the current AGV stops moving.

[0157] After the AGV completes its in-position movement toward the corresponding final support posture and stops, its actual parking position and actual vehicle direction may be completely consistent with the pre-set final support posture of the corresponding support position, or there may be differences due to positioning errors, motion control errors, or position deviations during the parking process.

[0158] First, based on the current AGV's actual parking position, actual vehicle direction, and vehicle outline information, determine the actual vehicle projection area after the current AGV stops moving.

[0159] Specifically, the actual parking position of the current AGV is used as the reference point position of its vehicle body. The orientation of the vehicle body outline is adjusted according to the actual vehicle body direction so that the vehicle body outline is in the posture defined by the actual parking position and the actual vehicle body direction, thereby obtaining the planar area actually covered by the current AGV in the support space.

[0160] The actual vehicle body projection area of ​​the current AGV is expanded outward according to the preset safety interval to form the actual occupancy envelope of the current AGV.

[0161] The actual occupancy envelope includes the actual vehicle body projection area covered after the current AGV stops moving, and the safety area located outside the actual vehicle body projection area.

[0162] The expansion method of the actual occupancy envelope is the same as that of the support occupancy envelope, both processed according to the same preset safety interval. The difference between the two is that the support occupancy envelope is generated based on the final support posture of the corresponding support position, while the actual occupancy envelope is generated based on the current AGV's actual parking position and actual vehicle orientation.

[0163] When the actual parking position and actual vehicle direction of the current AGV are consistent with the corresponding final support posture, the actual occupancy envelope of the current AGV is consistent with the support occupancy envelope of the corresponding support position.

[0164] When there is a difference between the actual parking position or actual vehicle orientation of the current AGV and the corresponding final support posture, the actual occupancy envelope of the current AGV will experience positional shift, directional deflection, or local coverage change relative to the support occupancy envelope of the corresponding support position.

[0165] After forming the actual occupancy envelope of the current AGV, determine the current support space state before the current AGV occupies its position.

[0166] The current support space state before the current AGV occupies its position is based on the initial support space state and includes the actual occupancy envelope of each AGV that completed its entry before the current AGV and is still within the support space.

[0167] For each prior AGV, the area corresponding to its actual occupancy envelope is determined as an impassable area and superimposed on the initial support space state.

[0168] When the current AGV is the first AGV to arrive at the corresponding final support posture, there is no AGV that has completed its entry before the current AGV. At this time, the current support space state before the current AGV occupies its position is consistent with the initial support space state.

[0169] When an AGV has already completed its support and occupancy, the actual occupancy envelope of the AGV is superimposed onto the initial support space state to form the current support space state before the current AGV occupies its position.

[0170] When multiple AGVs have already completed their support and occupying positions, the actual occupancy envelopes of each AGV are superimposed onto the initial support space state to form the current support space state, which includes the actual occupancy area of ​​each AGV.

[0171] Therefore, the current support space status reflects the actual occupancy of the passable area of ​​the support space by each preceding AGV that has been in the support space before the current AGV completes its occupation.

[0172] After the current support space state is formed, the support occupancy envelope of the current AGV's corresponding support position is superimposed on the current support space state as an impassable area to form a pre-calculated occupancy support space state.

[0173] The pre-calculated occupancy support space state indicates the support space state when each of the preceding AGVs maintains its position according to the actual parking state, and when the current AGV completes its occupancy according to the final support posture of the corresponding support position.

[0174] At the same time, the actual occupancy envelope of the current AGV is superimposed on the current support space state as an impassable area to form the actual occupancy support space state.

[0175] The actual occupancy support space state refers to the support space state when each preceding AGV maintains its position according to its actual parking state, and when the current AGV completes its occupancy according to its actual parking position and actual vehicle direction.

[0176] After the pre-calculated and actual occupancy support space states are established, the support positions that have not yet been installed are determined.

[0177] Support positions that have already been secured by prior AGVs are not considered support positions that have not yet been positioned. The support position corresponding to the current AGV is currently in the process of being secured and is not included in the determination of the terminal's approach to the channel's reachability.

[0178] For each of the remaining support positions that have not yet been positioned, determine the pre-calculated reachability of the corresponding terminal approach channel under the pre-calculated occupied support space state, and the actual reachability under the actual occupied support space state.

[0179] When determining the pre-calculated reachable state, the actual occupancy envelope of each preceding AGV and the support occupancy envelope of the current AGV's corresponding support position are deducted from the terminal approach channel to be judged, forming the remaining channel area under the pre-calculated occupancy state.

[0180] When determining the actual reachability, the actual occupancy envelopes of each preceding AGV and the current AGV are deducted from the terminal approach channel to be judged to form the remaining channel area under the actual occupancy status.

[0181] In the remaining channel area under the pre-calculated occupancy state and the remaining channel area under the actual occupancy state, a continuous posture sequence extending from the entry opening to the corresponding final support posture and satisfying the corresponding AGV motion constraint information is retrieved. When the continuous posture sequence is retrieved, the pre-calculated reachable state or the actual reachable state of the corresponding terminal approaching the channel is determined to be reachable. When the continuous posture sequence is not retrieved, the pre-calculated reachable state or the actual reachable state of the corresponding terminal approaching the channel is determined to be unreachable.

[0182] When the actual reachability of the approach channel for each support position that has not yet been completed is reached, it means that the current AGV's actual occupancy has not blocked the other AGVs from entering the continuous passage area of ​​the corresponding support position, thus confirming that the current AGV has completed its support occupancy.

[0183] Once the AGV has completed its support and positioning, its actual positioning envelope remains the non-passable area within the support space.

[0184] When there are still support positions that have not yet been filled, determine whether the preceding support position corresponding to each support position that has not yet been filled has been filled.

[0185] The support position that has completed the support occupation of all corresponding pre-positioning support positions is determined as the support position that meets the pre-positioning conditions, and the next support position to be inserted is determined from the support positions that meet the pre-positioning conditions.

[0186] When the actual reachability of the terminal corresponding to at least one unreachable support position is not yet completed, it means that after the current AGV completes its actual occupancy, at least one remaining AGV cannot continuously move from the entry opening to the corresponding final support posture, thus generating a channel blocking result.

[0187] The channel blocking result is determined based on the pre-calculated reachability and actual reachability of the blocked terminal approaching the channel.

[0188] When a terminal approaches a channel and its pre-calculated reachability is reachable, but its actual reachability is unreachable, the terminal's approach channel is identified as an actual occupancy deviation blocking channel.

[0189] The actual occupancy deviation blocking means that when the current AGV forms a theoretical occupancy according to the final support posture of the corresponding support position, the terminal approach channel is still reachable; after the current AGV forms an actual occupancy according to the actual parking position and the actual vehicle direction, the terminal approach channel becomes unreachable.

[0190] Therefore, the obstruction of the terminal approach channel is caused by the deviation of the current AGV's actual parking position or actual vehicle direction from the corresponding final support posture.

[0191] When both the pre-calculated reachability state and the actual reachability state of a terminal approaching a channel are unreachable, the terminal approaching channel is identified as a cumulative occupied blocking channel.

[0192] The cumulative occupancy blocking means that, even if the current AGV forms a support occupancy according to the corresponding final support posture, the terminal approach channel is still unreachable, provided that each of the preceding AGVs maintains its actual occupancy.

[0193] Therefore, the blockage of the terminal approach channel is caused by the actual occupancy of each preceding AGV and the current AGV's occupancy, and not just by the actual parking deviation of the current AGV.

[0194] When multiple terminal approach channels are in an unreachable state, the blocking type for each terminal approach channel is determined. The channel blocking result includes each blocked terminal approach channel and the blocking type corresponding to that terminal approach channel.

[0195] In a specific placement process of this embodiment, according to the aforementioned support placement dependency chain, the first AGV first reaches the final support posture corresponding to the first support position. The actual parking position and actual vehicle direction of the first AGV are basically consistent with the final support posture corresponding to the first support position. Its actual occupancy envelope does not make the terminal approach channels corresponding to the second, third, and fourth support positions unreachable. Therefore, it is confirmed that the first AGV has completed the support occupancy.

[0196] The fourth AGV then arrives at the area corresponding to the final support posture of the fourth support position and is identified as the current AGV. At this time, the first AGV is the preceding AGV, and the actual occupancy envelope of the first AGV is superimposed on the initial support space state to form the current support space state before the fourth AGV occupies its position.

[0197] After the support occupancy envelope corresponding to the fourth support position is superimposed on the current support space state, the terminal approach channels corresponding to the second and third support positions are both in a reachable state.

[0198] The actual parking position of the fourth AGV shifts towards the third support position relative to the final support posture corresponding to the fourth support position. Based on the actual parking position, actual vehicle direction, and vehicle outline information of the fourth AGV, the actual occupancy envelope occupies a portion of the continuous passage area in the terminal approach channel corresponding to the third support position.

[0199] After the actual occupancy envelope of the fourth AGV is superimposed, the terminal approach channel corresponding to the second support position can still retrieve a continuous attitude sequence that satisfies the motion constraint information of the second AGV, and its actual reachability is reachable.

[0200] The terminal approaching the third support position failed to retrieve a continuous attitude sequence that satisfies the motion constraint information of the third AGV, and its actual reachability is unreachable.

[0201] Since the pre-calculated reachability of the terminal approach channel corresponding to the third support position is reachable, but the actual reachability is unreachable, the terminal approach channel corresponding to the third support position is determined as the actual occupancy deviation blocking channel.

[0202] After generating the channel blocking result, perform channel restoration processing on the current AGV.

[0203] First, obtain the actual driving trajectory formed when the AGV enters the support space from the entry opening and moves to the actual parking position.

[0204] The actual driving trajectory is composed of multiple actual postures of the current AGV after entering the support space, arranged in the order of movement. Each actual posture includes the actual vehicle reference point position and actual vehicle direction of the current AGV at the corresponding location.

[0205] Starting from the current actual parking posture of the AGV, multiple candidate retreat postures are determined sequentially along the actual driving trajectory toward the direction of entering the opening.

[0206] The candidate retreat postures are all located on the actual driving trajectory of the current AGV, enabling the current AGV to move in reverse along the actual driving trajectory it has already passed to the corresponding candidate retreat posture.

[0207] The candidate retraction postures are distributed sequentially from near the actual parking position towards the entrance opening. Candidate retraction postures closer to the actual parking position correspond to smaller retraction distances, while candidate retraction postures closer to the entrance opening correspond to larger retraction distances.

[0208] For each candidate retreat posture, the vehicle body projection area of ​​the current AGV under that candidate retreat posture is determined based on the vehicle body reference point position, vehicle body direction, and vehicle body contour information when the current AGV is in that candidate retreat posture.

[0209] The vehicle body projection area is expanded outward according to the preset safety interval to obtain the retraction occupancy envelope of the corresponding candidate retraction posture.

[0210] For each candidate retreat posture, the current AGV actual occupancy envelope in the actual occupancy support space state is replaced with the retreat occupancy envelope corresponding to the candidate retreat posture.

[0211] The replacement refers to removing the actual occupancy envelope formed when the current AGV is in the actual parking position from the actual occupancy support space state, and superimposing the retreat occupancy envelope formed when the current AGV is in the candidate retreat posture in the same support space state.

[0212] Each replacement of the backoff placeholder envelope forms a backoff simulation support space state that matches the corresponding candidate backoff posture.

[0213] In each simulated backtracking support space state, the actual occupancy envelope of each preceding AGV remains unchanged, only the occupancy position and occupancy direction of the current AGV change.

[0214] For each backtracking simulation support space state, the reachability of each blocked terminal to the channel in the channel blocking results is reassessed.

[0215] The actual occupancy envelope of each preceding AGV and the retreat occupancy envelope corresponding to the current candidate retreat posture are deducted from the corresponding blocked terminal approach channel, and a continuous posture sequence that satisfies the corresponding AGV motion constraint information is retrieved in the remaining channel area.

[0216] If the continuous posture sequence is not found, it means that after the current AGV retreats to the candidate retreat posture, the blocked terminal approach channel has not yet been restored to a reachable state.

[0217] When the continuous posture sequence is retrieved, it means that after the current AGV retreats to the candidate retreat posture, the corresponding blocked terminal approach channel is restored to a reachable state.

[0218] When the channel blocking result includes multiple blocked terminals approaching the channel, the corresponding candidate backoff posture is determined as a candidate backoff posture that can release the channel only when all blocked terminals have recovered to the reachable state.

[0219] Among the candidate retreat postures that can restore each blocked terminal to a reachable state, the path distance between each candidate retreat posture and the current actual parking position of the AGV along the actual driving trajectory is determined.

[0220] The candidate backoff posture with the smallest path distance is selected as the channel release posture.

[0221] The channel release posture is the posture in which the AGV needs to retreat the shortest distance along its actual travel trajectory, provided that each blocked terminal can approach the channel and return to a reachable state.

[0222] In this embodiment, the first candidate retreat posture, the second candidate retreat posture, and the third candidate retreat posture are determined sequentially along the actual driving trajectory of the fourth AGV.

[0223] When the fourth AGV is in the first candidate retreat posture, the terminal approach channel corresponding to the third support position is still unreachable.

[0224] When the fourth AGV is in the second candidate retreat posture, it can retrieve a continuous posture sequence that satisfies the motion constraint information of the third AGV in the terminal approach channel corresponding to the third support position, and the terminal approach channel is restored to the reachable state.

[0225] When the fourth AGV is in the third candidate retreat posture, the terminal approach channel corresponding to the third support position is also in a reachable state.

[0226] Since the second candidate retraction posture is closer to the actual parking position of the fourth AGV than the third candidate retraction posture, the second candidate retraction posture is determined as the channel release posture, and the fourth AGV is controlled to move in the opposite direction along the actual driving trajectory to the channel release posture.

[0227] Once the AGV reaches the release posture of the channel, the support position corresponding to the current AGV will be redefined as a support position that has not yet been fully positioned.

[0228] Based on the actual position, actual vehicle direction, and vehicle outline information of the AGV when it is in the channel release posture, a corresponding retraction occupancy envelope is formed according to the preset safety interval.

[0229] The current AGV actual occupancy envelope in the actual occupancy support space state is replaced with the back-off occupancy envelope to form the support space state after the channel is restored.

[0230] The support space state after the channel is restored includes the actual occupancy envelope formed when each of the preceding AGVs maintains its support position, and the retreat occupancy envelope formed when the current AGV maintains its channel release posture.

[0231] During the specific positioning process, the first AGV remains in the first support position, and its actual occupancy envelope is retained in the support space state after the channel is restored; the fourth AGV remains in the channel release posture, and its retreat occupancy envelope replaces the original actual occupancy envelope.

[0232] At this point, the support levels that have not yet been fully entered include the second support level, the third support level, and the fourth support level, which has been redefined as not yet fully entered.

[0233] For each support position that has not yet been fully positioned, the support occupancy envelope corresponding to that support position is simulated in the support space state after the channel is restored.

[0234] For the second support position, the support occupancy envelope corresponding to the second support position is superimposed as an impassable area onto the support space state after the channel is restored, and the reachability state of the terminal approaching the channel corresponding to the third and fourth support positions before and after superposition is determined.

[0235] For the third support position, the support occupancy envelope corresponding to the third support position is superimposed as an impassable area onto the support space state after the channel is restored, and the reachability state of the terminal approaching the channel before and after superposition is determined for the second and fourth support positions.

[0236] For the fourth support position, since the support space state after the channel is restored already includes the retreat occupancy envelope formed when the fourth AGV is in the channel release posture, the retreat occupancy envelope of the fourth AGV is removed from the support space state after the channel is restored before simulating the support occupancy envelope corresponding to the fourth support position.

[0237] After removing the back-off placeholder envelope, the support placeholder envelope corresponding to the fourth support position is superimposed onto the corresponding support space state, and the reachability state of the terminal approach channel corresponding to the second and third support positions before and after superposition is determined.

[0238] By first removing the back-off placeholder envelope and then simulating the support placeholder envelope of the current AGV's corresponding support position, the same AGV corresponds to only one placeholder envelope in a simulated placeholder state.

[0239] For each simulated occupancy, the simulated superimposed support occupancy envelope and the actual occupancy envelope or retreat occupancy envelope already existing in the current state are deducted from the terminal approach channel to be judged, and a continuous posture sequence that satisfies the corresponding AGV motion constraint information is retrieved in the remaining channel area after occupancy.

[0240] When a support occupancy envelope corresponding to a support position that has not yet been filled causes the terminal approaching the channel of another support position that has not yet been filled to change from an reachable state to an unreachable state, a precedence relationship is established between the other support position that has not yet been filled and the current simulated support position.

[0241] In this embodiment, the support occupancy envelope corresponding to the second support position does not cause the terminal approach channel corresponding to the third and fourth support positions to change from an reachable state to an unreachable state.

[0242] The support occupancy envelope corresponding to the third support position causes the terminal approach channel corresponding to the fourth support position to change from an reachable state to an unreachable state. Therefore, the position precedence relationship between the fourth support position and the third support position is re-established.

[0243] The support occupancy envelope corresponding to the fourth support position did not cause the terminal approach channel corresponding to the second and third support positions to change from an accessible state to an inaccessible state.

[0244] Therefore, among the second, third, and fourth support levels that have not yet been filled, the fourth support level is the previous support level corresponding to the third support level; the second support level does not have a previous support level that has not yet been filled; and the fourth support level also does not have a previous support level that has not yet been filled.

[0245] Based on the newly determined positional relationship, the second, third, and fourth support positions are identified as support relationship nodes, and a directed connection is established from the support relationship node corresponding to the fourth support position to the support relationship node corresponding to the third support position.

[0246] The topology is layered according to the incoming connection status of each supporting node.

[0247] The support relationship nodes corresponding to the second and fourth support positions have no incoming connections, therefore, they are located in the regenerated first topology layer.

[0248] The support relationship node corresponding to the third support position has an incoming connection from the support relationship node corresponding to the fourth support position to that node, therefore, it is located in the regenerated second topology layer.

[0249] Therefore, in the regenerated support dependency chain, the second and fourth support positions are located in the first topology layer, and the third support position is located in the second topology layer.

[0250] Based on the newly generated support insertion dependency chain, control the corresponding AGV to continue executing the insertion.

[0251] Furthermore, referring to Figure 2, to achieve the above objectives, the present invention also provides an intelligent scheduling system for multi-AGV collaborative handling, the system comprising: The support space state construction module is used to acquire the spatial boundary information of the support space at the bottom of the target object, the entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transportation and corresponding to each support position, the vehicle body contour information and motion constraint information of each AGV, and the preset safety interval, and construct an initial support space state including a passable area based on the spatial boundary information, the entry opening and multiple support positions. The terminal approach channel generation module is used to generate a terminal approach channel that connects the entry opening to the corresponding final support posture for each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position. The occupancy relationship processing module is used to generate a support occupancy envelope for any of the support positions based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture, superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state, and determine the reachability of the terminal approach channels corresponding to the other support positions before and after the support occupancy envelope is superimposed; it is also used to establish an entry-precedence relationship between the second support position and the first support position when the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from a reachable state to an inaccessible state. The in-position dependency processing module is used to generate a support in-position dependency chain based on the in-position precondition relationship between each of the support positions. The positioning control module is used to control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support positioning dependency chain.

[0252] Specifically, the support space state construction module constructs the initial support space state according to step S1; the terminal approach channel generation module generates the terminal approach channel corresponding to each support position according to step S2; the occupancy relationship processing module generates the support occupancy envelope, forms the simulated occupancy support space state, and establishes the entry precondition relationship according to step S3; the entry dependency processing module generates the support entry dependency chain according to step S4; and the entry control module controls each AGV to reach the corresponding final support posture along the corresponding terminal approach channel according to the support entry dependency chain. The specific data processing process of each module is the same as the corresponding steps in Embodiment 1, and will not be repeated here.

[0253] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An intelligent scheduling method for collaborative handling of multiple AGVs, characterized in that, include: S1. Obtain the spatial boundary information of the bottom support space of the target object, the entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transportation and corresponding to each support position, the vehicle outline information and motion constraint information of each AGV, and the preset safety interval, and construct an initial support space state including a passable area based on the spatial boundary information, the entry opening and multiple support positions. S2. For each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position, generate a terminal approach channel that connects the entry opening to the corresponding final support posture. S3. For any of the support positions, generate a support occupancy envelope based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture. Superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state. Determine the reachability of the terminal approach channel corresponding to the other support positions before and after the support occupancy envelope is superimposed. When the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from an accessible state to an inaccessible state, an in-position precedence relationship is established between the second support position and the first support position. S4. Generate a support entry dependency chain based on the entry prerequisite relationship between each support position, and control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support entry dependency chain.

2. The intelligent scheduling method for multi-AGV collaborative handling according to claim 1, characterized in that, In step S2, the generation of the terminal proximity channel includes: Using the final support posture of the corresponding support position as the starting state for reverse analysis, multiple candidate posture nodes are generated towards the entry opening based on the vehicle body contour information and motion constraint information of the corresponding AGV. Based on the connection relationship between adjacent candidate posture nodes verified by motion constraints, a feasible posture sequence connecting the entry opening and the final support posture is generated. Based on the vehicle body sweeping area formed by the corresponding AGV moving along the feasible posture sequence, the vehicle body sweeping area is expanded outward according to the preset safety interval to obtain the terminal approach channel of the corresponding support position.

3. The intelligent scheduling method for multi-AGV collaborative handling according to claim 1, characterized in that, In step S3, the generation of the support occupancy envelope includes: Based on the final support posture of the corresponding support position and the vehicle body contour information of the corresponding AGV, determine the vehicle body projection area of ​​the AGV when it is in the final support posture; The vehicle body projection area is expanded outward according to the preset safety interval to obtain the support occupancy envelope of the corresponding support position.

4. The intelligent scheduling method for multi-AGV collaborative handling according to claim 3, characterized in that, In step S3, the reachability state of the terminal approaching the channel is determined in the following way: The impassable area corresponding to the support occupancy envelope is deducted from the terminal approach channel to be judged to form the remaining channel area after occupancy; Search within the remaining channel area after the occupancy of the space; retrieve a continuous posture sequence that extends from the entry opening to the corresponding final support posture and satisfies the corresponding AGV motion constraint information. When the continuous attitude sequence is retrieved, it is determined that the corresponding terminal is in a reachable state when approaching the channel; If the continuous attitude sequence is not found, it is determined that the corresponding terminal approach channel is in an unreachable state.

5. The intelligent scheduling method for multi-AGV collaborative handling according to claim 4, characterized in that, In step S4, the generation of the supporting in-position dependency chain includes: Each of the aforementioned support positions is determined as a support relationship node. Based on each of the aforementioned preceding support positions, a directed connection is established from the support relationship node corresponding to the preceding support position to the support relationship node corresponding to the following support position, forming a directed support entry relationship. The preceding support position corresponding to each support position is then determined. When there is no closed dependency path in the support-in directed relationship, the support relationship nodes are topologically layered according to the incoming connection state of each support relationship node, and the support-in dependency chain is generated according to the formation order of each topological layer.

6. The intelligent scheduling method for multi-AGV collaborative handling according to claim 5, characterized in that, The method further includes: S5. During the placement process according to the support placement dependency chain, the AGV that has reached the corresponding final support posture is determined as the current AGV, and the actual occupancy envelope of the current AGV is generated according to the actual parking position, actual vehicle direction and vehicle outline information of the current AGV and the preset safety interval. Based on the actual occupancy envelope of each AGV that completed its entry and remained in the support space before the current AGV occupies its position, the current support space state before the current AGV occupies its position is formed. The support occupancy envelope of the current AGV and the actual occupancy envelope of the current AGV are superimposed on the current support space state to form the pre-calculated occupancy support space state and the actual occupancy support space state. The pre-calculated reachability state and the actual reachability state of the terminal approach channel corresponding to each support position that has not yet been placed are determined respectively. When the actual reachability of each terminal approaching the channel is reachable, it is confirmed that the current AGV has completed the support position. If there are still support positions that have not been completed, the next support position to be entered is determined from the support positions that have been completed and occupied by the corresponding previous support positions. When at least one of the terminals approaches a channel whose actual reachability is unreachable, a channel blocking result is generated, channel recovery processing is performed on the current AGV, and the unexecuted part of the support placement dependency chain is updated according to the support space status after channel recovery.

7. The intelligent scheduling method for multi-AGV collaborative handling according to claim 6, characterized in that, The channel blocking result is generated based on the pre-calculated reachable state and the actual reachable state, wherein: The terminal approach channel that is pre-calculated as reachable but is actually unreachable is identified as the actual occupancy deviation blocking channel. Terminal approach channels that are unreachable in both the pre-calculated reachable state and the actual reachable state are identified as cumulative occupancy blocking channels. The channel blocking results include the blocked terminal approaching the channel and its corresponding blocking type.

8. The intelligent scheduling method for multi-AGV collaborative handling according to claim 7, characterized in that, The channel recovery process includes: The actual driving trajectory formed when the current AGV enters the support space is obtained, and multiple candidate retreat postures are determined along the actual driving trajectory; Based on the vehicle body contour information of the current AGV in each candidate retraction posture, a retraction occupancy envelope is generated, and the actual occupancy envelope in the actual occupancy support space state is replaced with each of the retraction occupancy envelopes to obtain multiple retraction simulated support space states. From the candidate retreat postures that restore the access channel of each blocked terminal to a reachable state, the candidate retreat posture with the smallest path distance along the actual driving trajectory from the actual parking position is selected as the channel release posture, and the current AGV is controlled to move to the channel release posture.

9. The intelligent scheduling method for multi-AGV collaborative handling according to claim 8, characterized in that, The updates to the unexecuted portions of the supporting in-place dependency chain include: The support position corresponding to the current AGV is redefined as the support position that has not yet been placed, and the actual occupancy envelope in the actual occupancy support space state is replaced with the retreat occupancy envelope when the current AGV is in the channel release posture, to form the support space state after the channel is restored. For each support position that has not yet been placed, the support occupancy envelope corresponding to the support position is simulated in the support space state after the channel is restored, and the influence of the support occupancy envelope on the terminal approaching the channel reachable state of the other support positions that have not yet been placed is determined. When simulating the support position corresponding to the current AGV, the retreat occupancy envelope is first removed from the support space state after the channel is restored. Based on the change in the reachable state, the pre-positioning relationship between each support position that has not yet been positioned, the corresponding pre-positioning support position and the support positioning dependency chain are re-determined, and the corresponding AGV is controlled to continue to perform the positioning according to the regenerated support positioning dependency chain.

10. An intelligent scheduling system for multi-AGV collaborative handling, characterized in that, include: The support space state construction module is used to acquire the spatial boundary information of the support space at the bottom of the target object, the entry opening connected to the support space, multiple support positions set in the support space, the final support posture corresponding to each support position, multiple AGVs participating in collaborative transportation and corresponding to each support position, the vehicle body contour information and motion constraint information of each AGV, and the preset safety interval, and construct an initial support space state including a passable area based on the spatial boundary information, the entry opening and multiple support positions. The terminal approach channel generation module is used to generate a terminal approach channel that connects the entry opening to the corresponding final support posture for each of the support positions, based on the initial support space state, the corresponding final support posture, and the vehicle body contour information and motion constraint information of the AGV corresponding to the support position. The occupancy relationship processing module is used to generate a support occupancy envelope for any of the support positions based on the vehicle body contour information of the AGV corresponding to the support position when it is in the corresponding final support posture, and to superimpose the area corresponding to the support occupancy envelope as an impassable area onto the initial support space state to form a simulated occupancy support space state, and to determine the reachability of the terminal approach channel corresponding to the other support positions before and after the support occupancy envelope is superimposed. It is also used to establish an in-position precedence relationship between the second support position and the first support position when the support occupancy envelope corresponding to the first support position causes the terminal approach channel corresponding to the second support position to change from an accessible state to an inaccessible state. The in-position dependency processing module is used to generate a support in-position dependency chain based on the in-position precondition relationship between each of the support positions. The positioning control module is used to control each AGV to enter the support space along the corresponding terminal approach channel and reach the corresponding final support posture according to the support positioning dependency chain.

Images