Unmanned forklift path planning method and device, and electronic equipment

Abstract

The present disclosure relates to the technical field of vehicle path planning, and provides an unmanned forklift path planning method, device and electronic equipment, which comprises: acquiring a topological road network; performing global path planning on the topological road network to generate an ordered node sequence; traversing adjacent node pairs of the ordered node sequence, and determining corresponding target driving modes of the forklift on each target road section according to a start point priority advancing strategy and a direction priority inheritance strategy; if the target driving modes corresponding to two adjacent target road sections are different, a turning candidate point is determined; at the turning candidate point, a target extension distance is extended along a target extension direction to determine a target turning point; and based on the ordered node sequence, the target road section, the target driving mode, the turning candidate point and the target turning point, a target motion path of the forklift is determined. The present scheme can adaptively determine an optimal turning point position, reduce unnecessary turning frequency of the forklift, and improve the moving distance proportion of the advancing motion mode of the forklift.

Description

Technical Field

[0001] This disclosure relates to the field of path planning and motion control technology for automated guided vehicles and autonomous mobile robots, specifically to a path planning method, device, and electronic equipment for unmanned forklifts. Background Technology

[0002] In warehousing and logistics and smart factory scenarios, unmanned forklifts need to frequently perform picking and placing tasks in structured environments such as platforms, aisles, and truck bodies. Unmanned forklifts have a clear heading, and their forward and backward movements are two different modes: when moving forward, the main detection direction of the unmanned forklift's perception system is wide and the maximum speed is high; when moving backward, the unmanned forklift's field of vision is limited and the maximum speed is low.

[0003] During the journey from the starting point to the destination, an unmanned forklift often cannot maintain a single movement mode. It typically needs to switch from forward to reverse or vice versa at certain points along the path. These points where movement modes switch can be called reversal points.

[0004] The location of the reversing point directly affects the driving efficiency, safety, and smoothness of the unmanned forklift. On the one hand, a reasonable selection of the reversing point can improve... high The proportion of forward travel distance is used to shorten the overall running time; on the other hand, the reversing point must provide sufficient reversing buffer space for the forklift. If the reversing point is close to the road node, the unmanned forklift will start in the opposite direction immediately after reaching the node, resulting in a harsh turning trajectory and large speed fluctuations, which is not conducive to the stability of the forklift's mechanical structure and the transported goods. Summary of the Invention

[0005] In view of this, one or more embodiments of this disclosure provide an unmanned forklift path planning method, device and electronic device, which can not only automatically plan a reasonable movement path from the start point to the end point for the unmanned forklift, but also adaptively determine the preferred reversing point position, thereby reducing the frequency of unnecessary reversing of the forklift and increasing the proportion of the moving distance in the forklift's forward movement mode.

[0006] Firstly, this disclosure provides a path planning method for an unmanned forklift, the method comprising: acquiring a topological road network composed of nodes and roads, wherein each segment of the road has a driving direction attribute, the driving direction attribute being used to constrain the heading of the target forklift; executing a global path planning algorithm on the topological road network according to the target start point and target end point to generate an ordered node sequence, wherein adjacent nodes of the ordered node sequence are connected by a target road segment; traversing each pair of adjacent nodes of the ordered node sequence, and determining the target driving mode corresponding to the target forklift on each target road segment according to the start point priority forward and direction priority inheritance strategy; wherein the target driving mode includes a forward mode and a backward mode; in the forward mode, the movement direction of the target forklift is consistent with the heading; in the backward mode, the movement direction of the target forklift is consistent with the heading. The heads are facing opposite directions; if the target driving modes corresponding to two adjacent target road segments are different, then the common nodes of the two adjacent target road segments are determined as candidate points for reversing direction; at the candidate points for reversing direction, the target extension distance is extended along a target extension direction to determine the target reversing point, wherein the target extension direction is different from the road direction of the two target road segments connecting the candidate points for reversing direction; wherein, the target reversing point is used for the target forklift to move to the candidate point for reversing direction after arriving at the candidate point along the previous target road segment, and then switch between forward and reverse modes, and enter the next target road segment through the candidate point for reversing direction according to the switched driving mode; based on the ordered node sequence, the target road segment, the target driving mode, the candidate points for reversing direction, and the target reversing point, the target movement path of the target forklift is determined.

[0007] Secondly, this disclosure provides an unmanned forklift path planning device, the device comprising: a road network information acquisition unit, used to acquire a topological road network composed of nodes and roads, wherein each road segment has a driving direction attribute, the driving direction attribute being used to constrain the heading of the target forklift; a path node planning unit, used to execute a global path planning algorithm on the topological road network according to the target start point and target end point, generating an ordered node sequence, wherein adjacent nodes of the ordered node sequence are connected by a target road segment; and a driving mode determination unit, used to traverse each pair of adjacent nodes of the ordered node sequence, and determine the target driving mode corresponding to the target forklift on each target road segment according to the start point priority forward and direction priority inheritance strategy; wherein the target driving mode includes a forward mode and a reverse mode; in the forward mode, the movement direction of the target forklift is consistent with the heading; in the reverse mode, the movement direction of the target forklift is consistent with the heading. The forklifts face opposite directions; a reversing point identification unit is used to identify common nodes of two adjacent target road segments as reversing candidate points if the target driving modes corresponding to the two adjacent target road segments are different; a reversing point extension unit is used to extend the target extension distance along a target extension direction at the reversing candidate point to determine the target reversing point, wherein the target extension direction is different from the road direction of the two target road segments connecting the reversing candidate point; wherein the target reversing point is used for the target forklift to move to the target reversing candidate point after arriving at the reversing candidate point along the previous target road segment, and then switch between forward and reverse modes, and enter the next target road segment through the reversing candidate point according to the switched driving mode; a motion path generation unit is used to determine the target motion path of the target forklift based on the ordered node sequence, the target road segment, the target driving mode, the reversing candidate point, and the target reversing point.

[0008] Thirdly, this disclosure provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, and when the computer program is executed by the processor, it implements the above-described unmanned forklift path planning method.

[0009] Fourthly, this disclosure provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the aforementioned unmanned forklift path planning method.

[0010] This disclosure provides an unmanned forklift path planning method in one or more embodiments. Through a global path planning algorithm, it generates an ordered sequence of nodes covering the target start and end points, locking in the preferred nodes and road segments that the forklift needs to pass through sequentially. Based on this, according to the driving direction attributes of each road segment, and combining a start-priority forward movement strategy and a direction-priority inheritance strategy, it can automatically determine whether the forklift's target driving mode on each road segment is forward or backward on the topological road network. This reduces unnecessary mode switching frequency and increases the proportion of forward travel distance. At locations where a necessary change in driving mode is required, a target reversing point can be adaptively formed by extending a distance along an extension direction. This makes the transition between forward and backward modes smoother, and eliminates the need to pre-mark fixed reversing points for each work location. Attached Figure Description

[0011] The features and advantages of the embodiments of this disclosure will be more clearly understood by referring to the accompanying drawings, which are illustrative and should not be construed as limiting the present disclosure in any way. In the drawings: Figure 1 A schematic diagram illustrating the steps of an unmanned forklift path planning method in one embodiment of this disclosure is shown. Figure 2 A flowchart illustrating an unmanned forklift path planning method according to one embodiment of this disclosure is shown. Figure 3 A schematic diagram illustrating the determination of a target driving mode and a target reversal point is shown in one embodiment of this disclosure; Figure 4 Another schematic diagram showing the determination of the target driving mode and the target reversal point in one embodiment of this disclosure is shown; Figure 5 A schematic diagram of the functional modules of an unmanned forklift path planning device in one embodiment of the present disclosure is shown. Figure 6 A schematic diagram of the structure of an electronic device according to one embodiment of the present disclosure is shown. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0013] In related technologies, unmanned forklifts typically employ a fixed reversing point scheme. For each work location, a fixed reversing point is pre-defined on the map; these reversing points are also called forward points or pre-stopping points. Map editors can manually mark the coordinates of the reversing point corresponding to each location on the map, ensuring that the unmanned forklift switches movement modes according to the preset fixed reversing points when performing tasks. This fixed reversing point scheme has the following drawbacks: First, the number of reversing points increases linearly with the number of storage locations, leading to high map maintenance costs. If the warehouse layout is adjusted (e.g., adding shelves, changing storage location numbers), the reversing points need to be modified one by one. When infeasible areas change, such as requiring temporary storage of goods or equipment relocation, the preset reversing points may fall into the new infeasible areas, preventing the forklift from completing the reversal upon arrival. Second, it cannot dynamically optimize for the specific start and end points of each task. Different loading and unloading tasks generally have different start and end points, but often share the same aisle section. Using fixed reversing points selects the same reversing position for all loading and unloading tasks, failing to optimize the forward travel distance for each task. During certain tasks, due to the unfavorable location of the fixed reversing point, the unmanned forklift may be forced to switch motion modes multiple times, affecting the forklift's operating efficiency.

[0014] In related technologies, some solutions utilize global path planning algorithms, such as A Search algorithms, such as Dijkstra's algorithm, can plan a complete path for a forklift from its starting point to its destination and directly select a road node as a reversal point along that path. However, the reversal points selected by these methods generally lack buffer zones, causing the forklift to immediately reverse and start again at the road node, resulting in abrupt turns. This is especially noticeable in narrow aisles or inside a truck bed, where frequent sudden stops and starts can adversely affect the stability of the goods and the mechanical structure of the forklift.

[0015] In view of this, the unmanned forklift path planning method provided by one or more embodiments of this disclosure can automatically determine the target driving mode of the forklift on each road segment based on the driving direction attribute of the road, thereby increasing the proportion of forward driving distance and reducing the frequency of unnecessary mode switching. At locations where a necessary change in driving mode is required, a target reversing point is formed by extending a distance along an extension direction. This makes the transition between forward and reverse modes of the forklift smoother and eliminates the need to pre-mark fixed reversing points for each work location.

[0016] Please see Figure 1 The unmanned forklift path planning method provided in one embodiment of this disclosure may include the following steps S1-S6. The unmanned forklift path planning method provided in this embodiment can be executed by the local computing module of the unmanned forklift or by the remote control module of the unmanned forklift.

[0017] S1: Obtain the topological road network composed of nodes and roads, wherein each road segment has a driving direction attribute, and the driving direction attribute is used to constrain the direction of the target forklift.

[0018] In this implementation, nodes represent key locations in the road network. Nodes can possess attributes such as node coordinates, node orientation, node name, and names of preceding nodes. Roads can connect different nodes and can possess attributes such as road start point, road end point, road length, maximum permissible speed, and driving direction. The driving direction attribute specifies the permissible orientation of the forklift's front end when the unmanned forklift is traveling on this road.

[0019] In this embodiment, the driving direction attribute includes a forward attribute, a reverse attribute, and a bidirectional attribute. The forward attribute is used to limit the direction of the target forklift's front end to a first direction, the reverse attribute is used to limit the direction of the target forklift's front end to a second direction, and the bidirectional attribute is used to allow the target forklift's front end to face either the first or the second direction. The first direction points from the road start node to the road end node, and the second direction points from the road end node to the road start node. The driving direction attribute does not restrict the target forklift from entering the road from either end. The driving direction attribute is used to determine the forward or backward mode of the target forklift on the road by constraining the direction of the target forklift's front end.

[0020] Specifically, the driving direction attribute mainly includes forward, reverse, and bidirectional. When the driving direction attribute is forward, the forklift must be facing the direction from the start point to the end point of the road when traveling on the corresponding road. When the driving direction attribute is reverse, the forklift must be facing the direction from the end point to the start point of the road when traveling on the corresponding road. When the driving direction attribute is bidirectional, the forklift can be facing either the direction from the start point to the end point of the road or the direction from the end point to the start point of the road when traveling on the corresponding road.

[0021] It's important to note that the driving direction attribute constrains the forklift's heading, not whether the target forklift can enter the road from one end. Except for restricted areas, roads with any driving direction attribute generally allow forklifts to enter from either end; the only difference is whether the forklift should move forward or backward upon entering the road. In other words, subsequent global path planning algorithms can generate an ordered sequence of nodes without considering the driving direction attribute, only needing to consider the connectivity between adjacent ordered nodes.

[0022] S2: Based on the target starting point and target ending point, execute a global path planning algorithm on the topology network to generate an ordered node sequence, wherein adjacent nodes in the ordered node sequence are connected by a target road segment.

[0023] In this embodiment, based on the specific loading and unloading tasks to be performed each time, the target starting point S and the target ending point T can be clearly defined on the topology network. A global path planning algorithm, such as A..., is then executed on the topology network. The Dijkstra algorithm can automatically search for a reasonable path from the starting point S to the ending point T. This reasonable path can be represented by an ordered sequence of nodes. It means that, among them , .

[0024] In this embodiment, each node in the ordered node sequence is a road node that the forklift needs to pass through sequentially. Adjacent nodes in the ordered node sequence are connected by a road segment. The ordered node sequence mainly determines the topological orientation of the forklift's movement path, that is, which nodes the forklift passes through, and does not involve determining the specific movement mode of the forklift, such as forward or backward.

[0025] S3: Traverse each adjacent node pair of the ordered node sequence, and determine the target driving mode of the target forklift on each target road segment according to the starting point priority and direction priority inheritance strategy; wherein, the target driving mode includes forward mode and backward mode; in forward mode, the movement direction of the target forklift is consistent with the direction of the vehicle head; in backward mode, the movement direction of the target forklift is opposite to the direction of the vehicle head.

[0026] In this implementation, the starting point-priority forward movement and direction-priority inheritance strategy ensures that starting from the initial node (i.e., the target starting point), the forklift prioritizes the forward movement mode, travels through the first target road segment, and arrives at the second node in the ordered node sequence. In subsequent target road segments, the forklift prioritizes maintaining the same driving mode as the previous target road segment. Only when encountering a one-way road with a driving direction attribute, and when the driving direction attribute conflicts with the driving mode inherited by the target forklift, will the forklift change its driving mode.

[0027] Specifically, for the first road segment, if the driving direction attribute allows the target forklift to travel in forward mode, then the corresponding target driving mode is determined to be forward mode; otherwise, the corresponding target driving mode is determined to be reverse mode. For any target road segment other than the first road segment, when the driving direction attribute of the target road segment is a single vehicle heading constraint: if the vehicle heading constraint conflicts with the driving mode inherited by the target forklift, then the target driving mode of the target forklift is determined to be the opposite driving mode to that on the previous target road segment; if the vehicle heading constraint does not conflict with the driving mode inherited by the target forklift, then the target driving mode of the target forklift is determined to be the same driving mode as that on the previous target road segment; when the driving direction attribute of the target road segment is a bidirectional attribute, the target forklift directly inherits the same driving mode as that on the previous target road segment.

[0028] Specifically, for an ordered sequence of nodes Each of the adjacent node pairs The number can be determined according to the following rules. section of road Connecting nodes and nodes Forklift driving mode .

[0029] For the first section of road (Right now If the driving direction attribute allows, the forklift is defaulted to forward mode; if... For one-way roads and based on their driving direction attributes from arrive If the forklift must move backward, then the forklift will be forcibly classified as being in reverse mode.

[0030] For the section of road ( Forklifts inherit the driving mode from the previous road segment by default. .

[0031] like The road is a one-way road and its driving direction attribute does not allow forklifts to be used in inheritance mode. Movement, that is, the actual movement of the forklift from... arrive If the direction of the vehicle conflicts with the road's defined frontal direction, then it will be forced to... Determined as The opposite direction; otherwise, keep .

[0032] It should be noted that for roads with a two-way driving direction attribute, since the driving direction attribute does not specifically limit the direction of the vehicle head, the forklift will necessarily meet the priority inheritance condition, so that the forklift can maintain the same driving mode as the previous road segment on two-way roads.

[0033] S4: If the target driving modes corresponding to two adjacent target road segments are different, then the common nodes of the two adjacent target road segments are determined as candidate points for reversal.

[0034] In this embodiment, when the target driving mode of the forklift changes between two adjacent road segments, the node where the change occurs can be identified as a candidate point for direction reversal. For example, when At that time, node This is the candidate point for reversing, meaning that the forklift needs to switch from forward mode to reverse mode, or from reverse mode to forward mode, at this point.

[0035] S5: At the candidate reversal point, extend the target extension distance along a target extension direction to determine the target reversal point. The target extension direction is different from the road direction of the two target road segments connecting the candidate reversal point. The target reversal point is used for the target forklift to move to the candidate reversal point after arriving at the candidate reversal point along the previous target road segment, and then switch between forward and reverse modes. According to the switched driving mode, it drives into the next target road segment through the candidate reversal point.

[0036] In this embodiment, the target reversing point is used to provide a reversing buffer space for the target forklift. After the target forklift travels along the previous target road segment to the reversing candidate point, it does not directly switch driving modes at the reversing candidate point. Instead, it continues to move the target extension distance along the target extension direction until it reaches the target reversing point. During the process of moving from the reversing candidate point to the target reversing point, the target forklift can gradually reduce its speed and adjust its body posture. After reaching the target reversing point, it switches between forward and reverse modes and enters the next target road segment through the reversing candidate point according to the switched driving mode.

[0037] Therefore, the target reversal point is equivalent to forming a buffer position near the candidate reversal point, enabling the target forklift to avoid immediately reversing at the road intersection, thereby improving the smoothness of the reversal process and the controllability of the path execution.

[0038] In some implementations, at a candidate reversal point, the target reversal point is determined by extending the target extension distance along a target extension direction. This includes: determining all adjacent road directions of the candidate reversal point based on the topological road network; excluding the road directions of two target road segments connected to the candidate reversal point from the adjacent road directions, and using the remaining road directions as candidate extension directions; calculating the angle between each candidate extension direction and the original direction of the forklift, denoted as the candidate angle, where the original direction of the forklift is the direction of travel of the target forklift before entering the candidate reversal point; sorting the candidate extension directions according to the angle difference between each candidate angle and a preset angle from smallest to largest, and determining the candidate extension direction that ranks first and meets the preset conditions as the target extension direction; wherein the preset conditions include at least one of collision verification, the extendable length meeting the target extension distance, and the road width meeting a preset width threshold.

[0039] Specifically, after excluding the "entering direction" already used to enter the candidate for reversing point and the "exit direction" about to leave the candidate for reversing point, the remaining adjacent road directions of the candidate for reversing point can be automatically used as candidate extension directions. Among the candidate extension directions, the direction with an angle close to the current direction of travel of the forklift is preferentially selected as the target extension direction, wherein the preset angle is preferably 90°, or the preset angle is within the range of 60° to 120° to ensure a smoother reversing trajectory.

[0040] In a practical application example, if there is no candidate extension direction that makes the candidate angle close to or equal to 90°, then the candidate extension direction that minimizes the deviation of the candidate angle from 90° is selected as the target extension direction.

[0041] In a practical application example, if there are multiple candidate extension directions that make the candidate included angle close to or equal to 90°, the candidate extension direction that passes the collision check, has an extendable length that meets the target extension distance, and has a road width that meets the preset width threshold is selected as the target extension direction.

[0042] In some implementations, at the candidate reversal point, the target reversal point is determined by extending the target extension distance along a target extension direction, including: The target extension distance is determined using the following formula; ; in, Extend the distance to the target. The speed at which the target forklift arrives at the candidate reversing point. The preset maximum permissible deceleration can be determined based on at least one of the following: cargo weight, cargo center of gravity height, and cargo friction coefficient. This is a preset distance constant.

[0043] Specifically, the target extension distance can be determined based on the operating status of the target forklift. If a fixed value is used for the target extension distance, when the target forklift enters the reversing candidate point at a high speed, with a large load, or with a high center of gravity of the goods, the fixed extension distance may not be sufficient to provide adequate deceleration buffer space, thereby increasing the risk of goods slipping, tipping over, or forklift posture fluctuations.

[0044] Therefore, in some implementations, the target extension distance can be determined based on the speed of the target forklift when it arrives at the candidate reversing point, a preset maximum allowable deceleration, and a preset distance constant. The higher the speed of the target forklift, or the smaller the preset maximum allowable deceleration, the greater the target extension distance. This allows the target forklift to have more deceleration and attitude adjustment space during its movement from the candidate reversing point to the target reversing point, improving the smoothness of the reversing process.

[0045] S6: Based on the ordered node sequence, the target road segment, the target driving mode, the candidate reversal point, and the target reversal point, determine the target movement path of the target forklift.

[0046] In this embodiment, the target movement path includes multiple target road segments connected sequentially according to an ordered node sequence, and extension segments inserted at candidate reversal points. For each target road segment, the target driving mode corresponding to the target road segment is recorded; for each candidate reversal point, the target extension direction, target extension distance, and target reversal point corresponding to the candidate reversal point are recorded.

[0047] When the target forklift executes the target movement path, the target forklift drives according to the target driving mode corresponding to each target road segment. When it reaches the candidate point for reversing direction, the target forklift first moves the target extension distance along the target extension direction to reach the target reversing point. At the target reversing point, it performs a switch between forward mode and reverse mode, and then drives into the next target road segment through the candidate point for reversing direction according to the switched target driving mode.

[0048] In some implementations, the unmanned forklift path planning method may further include a collision verification step before finally outputting the target motion path. This collision verification step specifically includes: obtaining no-entry indication information marking infeasible areas in the topological road network; based on the no-entry indication information, detecting whether the vehicle outline of the target forklift interferes with the infeasible area on the target motion path; if the vehicle outline does not interfere with the infeasible area, then the target motion path is confirmed to be usable; if the vehicle outline interferes with the infeasible area, and the interference is located in the extension segment from the candidate reversal point to the target reversal point, then the candidate extension directions are switched sequentially according to their order, and the target reversal point is re-determined; if all candidate extension directions cause interference, then the road direction or node direction combination causing the interference is masked, and the global path planning algorithm is re-executed.

[0049] Specifically, based on the no-entry instruction information, on the target movement path, it is detected whether the vehicle profile of the target forklift interferes with the infeasible area. This includes: determining multiple pose points along the complete target movement path with a fixed step size; at each pose point, determining the forklift's heading and vehicle profile based on the road direction of the target road segment where the pose point is located, the target driving mode corresponding to the target road segment, and the geometric dimensions of the target forklift; and checking whether the vehicle profile overlaps with the infeasible area at any pose point. The complete target movement path includes the target road segment corresponding to the ordered node sequence and the extension segment between the candidate reversal point and the target reversal point.

[0050] Every fixed distance The preferred sampling interval is 0.5m per pose point. Each pose point can contain coordinate information and vehicle heading information. At each sampling point, based on the forklift geometry, such as length... ,width and the direction the front of the car is facing First, draw the vehicle outline of the forklift. Then, check whether each vehicle outline overlaps with any infeasible region in the topological road network map. If the vehicle outlines of all sampled points do not interfere with any infeasible regions, the current target movement path is confirmed as feasible.

[0051] After determining the target extension direction and the target reversal point, this scheme also performs vehicle contour interference detection on the complete path, including the extension segment. If interference occurs in the extension segment, it prioritizes switching to other candidate extension directions; if all candidate extension directions are unavailable, it masks the road direction or node direction combination that causes the interference and re-executes the global path planning algorithm. Through the above processing method, unnecessary global replanning times can be reduced, and the executability of the target motion path can be improved.

[0052] Furthermore, in actual forklift operation, the goods being transported may be excessively long or wide, and the forklift forks may extend to varying degrees. Therefore, by using an onboard 3D vision camera or LiDAR to scan the actual outer contour of the goods in real time, the current forklift contour can be dynamically generated, and then collision verification can be performed.

[0053] Therefore, the output target motion path includes not only the nodes and road segments that the target forklift needs to pass through, but also the forward or backward mode of each road segment, candidate reversal points, target reversal points, and extension segment information, so that the motion control module of the target forklift can execute it segment by segment.

[0054] Please see Figure 2 This disclosure provides an implementation method for unmanned forklift path planning, the overall process of which is as follows: based on given start and end point information, first execute A on the topology network. The pathfinding algorithm generates an ordered sequence of nodes. Following the expansion direction of this ordered sequence, and adhering to the rules of starting point priority and direction priority inheritance, it determines the forklift's forward / backward movement pattern on each target road segment. At points where the forklift's movement pattern changes, candidate turning points are identified and extended along the target extension direction to form target turning points. Subsequently, collision checks based on the forklift's outline are performed on the complete target movement path. If the forklift outline collides with an infeasible area on the map at a candidate turning point's extension segment, other candidate extension directions are switched and the process is retried. If all candidate extension directions for a candidate turning point would cause interference, the turning action at that candidate turning point is deemed infeasible. In this case, the road directions or node direction combinations causing interference can be masked, and the global path planning algorithm is re-executed to generate a new ordered sequence of nodes that bypasses conflicting road directions or node direction combinations. The node direction combination can include at least two of the following: the forklift's inbound direction when entering a candidate turning point, its outbound direction when leaving a candidate turning point, and the candidate extension direction selected at that candidate turning point.

[0055] Please see Figure 3 In one embodiment of this disclosure, after an unmanned forklift retrieves goods from a picking point S deep within the warehouse, it needs to transport the goods to a drop-off point T at the end of the platform. There is a section between the warehouse and the platform... This is a special passage. The map defines this passage as a one-way, reverse road, meaning the direction the vehicle is facing is specified as... Direction, forklift from Towards When driving, you must drive in reverse mode.

[0056] The topological road network consisting of nodes and roads is The corresponding warehouse alleyway has a two-way traffic direction attribute. Figure 3 (Indicated by two dashed arrows pointing in opposite directions). The corresponding passageway connecting the warehouse and the platform has a two-way traffic direction attribute. Figure 3 (Indicated by two dashed arrows pointing in opposite directions). For special passages, the road's driving direction attribute is reverse ( Figure 3 (Illustrated by a dashed arrow pointing in a single direction). The corresponding platform end passageway has a two-way traffic direction attribute ( Figure 3 (The image is indicated by two dashed arrows pointing in different directions).

[0057] In this embodiment, the unmanned forklift path planning method may include the following process.

[0058] Process 1, A Finding a way. Starting point As the endpoint, execute A on the topology network. The algorithm yields an ordered sequence of nodes. .

[0059] Process 2, Directional Inheritance Determination. First Segment Based on the principle of prioritizing the starting point, the forklift's driving mode in this section is determined to be forward. Figure 3 (Indicated by a solid arrow in the middle), denoted as Second section If the forklift inherits the same driving mode as the previous road segment, no conflict will occur, and the forklift's driving mode in this road segment will be determined to be forward. Figure 3 (Indicated by a solid arrow in the middle), denoted as Third section If the forklift inherits the same driving mode as the previous road segment, a conflict event will occur, and the forklift's driving mode will be determined to change to reverse in this road segment. Figure 3 (Indicated by a solid arrow in the middle), denoted as Section 4 If the forklift inherits the same driving mode as the previous road segment, no conflict will occur, and the forklift's driving mode in this road segment will be determined to be reverse. Figure 3 (Indicated by a solid arrow in the middle), denoted as .

[0060] Step 3: Identify candidate reversal points. (Checked) , Forklift driving mode at node Changes are needed at this location; determine the node. These are candidate points for reversal.

[0061] Step 4: Extend the reversal point. For nodes... Its adjacent roads include (Forklift direction) (Forklift exit direction) and a side aisle After excluding the forklift's inbound and outbound directions, the candidate extension direction is determined as follows: Forklift from node ,along Direction, extension m, reaching the target reversal point.

[0062] Step 5: Collision Check. Along the complete target movement path (including extensions), sample the predicted pose of the forklift every 0.5 m, draw a 2.0 m × 1.0 m forklift outline rectangle, and perform interference checks with infeasible areas on the topology road network map.

[0063] Step 6: Execution. Assuming no interference at any sampling point, the forklift can adopt the target motion path that has successfully passed collision verification. Afterwards, the forklift can proceed from the node... First, proceed in forward driving mode, passing through nodes. Reaching the node ; at the node At the target reversing point, the forklift continues in forward mode, advancing 1.0 m along the target direction. At the reversing point, the forklift switches its mode from forward to reverse, and then continues along... The path is to retreat backwards to reach the destination. .

[0064] In this embodiment, the forklift is in This section (approximately 60% of the total path) travels at a relatively high forward speed of 1.5 m / s. The forklift only... The forklift was forced to travel in reverse at a speed of 0.5 m / s due to the rigid constraints of the road's directional properties. This segment inherits the reverse driving mode. Throughout the entire path, the forklift only operates at nodes... It changes direction once and uses a 1.0 m extension section to achieve a smooth transition between driving modes.

[0065] Please see Figure 4 In a practical application example of this implementation, if the node Nearby side alleyways The temporary storage of goods at this point creates an infeasible area, rendering the preferred extension direction at that point infeasible. Assume node... Except for the direction of entry ( ) and outward direction ( In addition to the above, there are two candidate extension directions: (The lateral aisle forms a 90° angle with the forklift's current direction of travel); (Another lateral aisle forms a 135° angle with the forklift's current direction of travel). After sorting the candidate extension directions by the absolute difference between the angle and the preferred preset angle of 90° from smallest to largest, the priority extension directions to try are: However, during the initial collision check, along the direction... The extended segment was used to draw a sampling rectangle of the forklift outline, and interference with infeasible areas (stacked goods) was found. Figure 4 middle, A solid arrow marked in red indicates that the collision check failed. (Not set as the target reversal point). Therefore, switch the extension direction according to the sorting, along the direction is... Draw the forklift outline sampling rectangle on the extended segment and re-perform collision verification. After no interference, the forklift... place, along Extend 1.0 m in the direction to reach the target reversal point. ( Figure 4 In the middle, the reversing point Collision check passed, set as target reversing point), switch to reverse driving, and finally follow... Reverse back to reach the finish line.

[0066] In this application example, when the preferred extension direction collides with an infeasible area, it can automatically switch to other candidate extension directions, avoiding directly entering a time-consuming rerouting process.

[0067] In a practical application example of this implementation, if the node All candidate extension directions interfere with the infeasible area, preventing the forklift from proceeding. The reversal is completed at this point. At this time, it can lead to... In conflict Section A Remove from the search graph and then re-execute A. Finding a way. After starting the search again, A The search algorithm found an alternative route. ,in This is another passable path. After repeating the direction inheritance determination and collision verification process for the new path, it is assumed that the new path is at a new candidate reversal point. The collision check passed, and the forklift can be instructed to travel along the new path. The reversal is completed at this point.

[0068] It needs to be further explained that, Figure 3 and Figure 4 In this context, the "solid arrow" only indicates the direction of the forklift's front when it moves, and does not limit the forklift's movement path.

[0069] This disclosure provides an unmanned forklift path planning method with one or more embodiments. Through a global path planning algorithm, it generates an ordered sequence of nodes covering the target start and end points, locking in the preferred nodes and road segments that the forklift needs to pass through sequentially. Based on this, according to the driving direction attributes of each road segment, and combining a start-priority forward movement strategy and a direction-priority inheritance strategy, it can automatically determine whether the forklift is moving forward or backward on each road segment within the topological road network, reducing unnecessary mode switching frequency and increasing the proportion of forward travel distance.

[0070] At locations where a necessary change in driving mode is required, a target reversing point is adaptively formed by extending a certain distance along an extension direction. In this scheme, the target reversing point is not directly set at the candidate reversing point, but rather further extended in the adjacent direction. The target forklift can decelerate and adjust its posture during this extension process, and switch driving modes at the target reversing point. Subsequently, the target forklift, following the switched driving mode, enters the next target road segment via the candidate reversing point. Through this method, the reversing action no longer occurs concentrated at road junctions, but rather forms a buffer zone using the extended space near the candidate reversing point. This helps reduce the impact of sudden stops, starts, or turns at junctions, improving the smoothness of the transition between forward and reverse modes.

[0071] Please see Figure 5 This disclosure also provides an unmanned forklift path planning device, the device comprising: The road network information acquisition unit 100 is used to acquire a topological road network composed of nodes and roads, wherein each road segment has a driving direction attribute, and the driving direction attribute is used to constrain the direction of the target forklift. The path node planning unit 200 is used to execute a global path planning algorithm on the topological road network according to the target starting point and the target ending point to generate an ordered node sequence, wherein adjacent nodes in the ordered node sequence are connected by a target road segment. The driving mode determination unit 300 is used to traverse each adjacent node pair of the ordered node sequence and determine the target driving mode of the target forklift on each of the target road segments according to the starting point priority forward and direction priority inheritance strategy; wherein, the target driving mode includes a forward mode and a reverse mode; in the forward mode, the movement direction of the target forklift is the same as the front of the vehicle; in the reverse mode, the movement direction of the target forklift is opposite to the front of the vehicle. The reversal point identification unit 400 is used to identify the common node of the two adjacent target road segments as a reversal candidate point if the target driving modes corresponding to the two adjacent target road segments are different. The reversing point extension unit 500 is used to extend a target extension distance along a target extension direction at the reversing candidate point to determine a target reversing point. The target extension direction is different from the road direction of the two target road segments connecting the reversing candidate point. The target reversing point is used for the target forklift to move to the target reversing point after arriving at the reversing candidate point along the previous target road segment, and then switch between forward mode and reverse mode. According to the switched driving mode, it drives into the next target road segment through the reversing candidate point. The motion path generation unit 600 is used to determine the target motion path of the target forklift based on the ordered node sequence, the target road segment, the target driving mode, the candidate reversal point, and the target reversal point.

[0072] In one embodiment, the driving direction attribute includes a forward attribute, a reverse attribute, and a bidirectional attribute; the forward attribute is used to limit the direction of the target forklift's front end to a first direction, the reverse attribute is used to limit the direction of the target forklift's front end to a second direction, and the bidirectional attribute is used to allow the target forklift's front end to face both the first and second directions; wherein, the first direction points from the road start node to the road end node, the second direction points from the road end node to the road start node, the driving direction attribute does not restrict the target forklift from entering the road from either end, and the driving direction attribute is used to constrain the direction of the target forklift's front end to determine the target forklift's forward or backward mode on the road.

[0073] In one embodiment, the driving mode determination unit 300 is specifically used for: for the first road segment, if the driving direction attribute allows the target forklift to travel in a forward mode, then the corresponding target driving mode is determined as a forward mode; otherwise, the corresponding target driving mode is determined as a reverse mode; for any target road segment other than the first road segment, when the driving direction attribute of the target road segment is a single vehicle heading constraint: if the vehicle heading constraint conflicts with the driving mode inherited by the target forklift, then the target driving mode of the target forklift is determined to be the opposite driving mode to that on the previous target road segment; if the vehicle heading constraint does not conflict with the driving mode inherited by the target forklift, then the target driving mode of the target forklift is determined to be the same driving mode as that on the previous target road segment; when the driving direction attribute of the target road segment is a bidirectional attribute, the target forklift directly inherits the same driving mode as that on the previous target road segment.

[0074] In one embodiment, the reversing point extension unit 500 is specifically used for: determining all adjacent road directions of the candidate reversing point according to the topological road network; excluding the road directions of the two target road segments connected to the candidate reversing point from the adjacent road directions, and taking the remaining road directions as candidate extension directions; calculating the angle between each candidate extension direction and the original direction of the forklift, and recording it as the candidate angle, where the original direction of the forklift is the driving direction of the target forklift before entering the candidate reversing point; sorting the candidate extension directions according to the angle difference between each candidate angle and a preset angle from small to large, and determining the candidate extension direction that ranks first and meets the preset conditions as the target extension direction; wherein the preset conditions include at least one of collision verification, the extendable length meeting the target extension distance, and the road width meeting a preset width threshold.

[0075] In one embodiment, the reversing point extension unit 500 is specifically used to: determine the target extension distance using the following formula; ; in, Extend the distance to the target. The speed at which the target forklift arrives at the candidate reversal point. The preset maximum permissible deceleration is determined based on at least one of the following: cargo weight, cargo center of gravity height, and cargo friction coefficient. This is a preset distance constant.

[0076] In one embodiment, the unmanned forklift path planning device further includes a collision detection unit, which is used to: acquire no-entry indication information marking infeasible areas in the topological road network; detect whether the vehicle outline of the target forklift interferes with the infeasible area on the target movement path according to the no-entry indication information; if the vehicle outline does not interfere with the infeasible area, then the target movement path is confirmed to be usable; if the vehicle outline interferes with the infeasible area, and the interference is located in the extension segment from the candidate reversal point to the target reversal point, then the candidate extension directions are switched sequentially according to the order of the candidate extension directions, and the target reversal point is re-determined; if all candidate extension directions cause interference, then the road direction or node direction combination causing the interference is masked, and the global path planning algorithm is re-executed.

[0077] In one embodiment, the collision detection unit is specifically used to: determine multiple pose points along the complete path of the target motion path at fixed step sizes; at each pose point, determine the frontal orientation and vehicle profile of the target forklift based on the road direction of the target road segment where the pose point is located, the target driving mode corresponding to the target road segment, and the geometric dimensions of the target forklift; check whether the vehicle profile overlaps with the infeasible area at any pose point; wherein, the complete path includes the target road segment corresponding to the ordered node sequence and the extension segment between the candidate reversal point and the target reversal point.

[0078] The various units described in the above embodiments can be implemented by computer chips or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a smartphone, a personal digital assistant, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0079] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0080] Please see Figure 6 This disclosure also provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, and when the computer program is executed by the processor, it implements the above-described unmanned forklift path planning method.

[0081] This disclosure also provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the above-described unmanned forklift path planning method.

[0082] The processor can be a central processing unit (CPU). It can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof.

[0083] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby implementing the methods in the above-described embodiments.

[0084] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0085] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0086] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, embodiments of apparatus, devices, and storage media are basically similar to method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0087] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

[0088] Although embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A path planning method for an unmanned forklift, characterized in that, The method includes: Obtain a topological road network consisting of nodes and roads, wherein each road segment has a driving direction attribute, and the driving direction attribute is used to constrain the heading of the target forklift; Based on the target start point and target end point, a global path planning algorithm is executed on the topological road network to generate an ordered node sequence, wherein adjacent nodes in the ordered node sequence are connected by a target road segment; Traverse each adjacent node pair of the ordered node sequence, and determine the target driving mode of the target forklift on each target road segment according to the starting point priority and direction priority inheritance strategy; wherein, the target driving mode includes a forward mode and a reverse mode; in the forward mode, the movement direction of the target forklift is the same as the direction of the vehicle head; in the reverse mode, the movement direction of the target forklift is opposite to the direction of the vehicle head. If the target driving modes corresponding to two adjacent target road segments are different, then the common nodes of the two adjacent target road segments are determined as candidate points for reversal. At the candidate turning point, the target turning point is determined by extending the target extension distance along a target extension direction. The target extension direction is different from the road direction of the two target road segments connecting the candidate turning point. The target turning point is used for the target forklift to move to the candidate turning point after arriving at the candidate turning point along the previous target road segment, and then switch between forward and reverse modes. According to the switched driving mode, it drives into the next target road segment through the candidate turning point. Based on the ordered node sequence, the target road segment, the target driving mode, the candidate turning points, and the target turning points, the target movement path of the target forklift is determined. The step of determining the target reversing point by extending the target extension distance along a target extension direction at the candidate reversing point includes: determining all adjacent road directions of the candidate reversing point according to the topological road network; excluding the road directions of the two target road segments connected to the candidate reversing point from the adjacent road directions, and taking the remaining road directions as candidate extension directions; determining the angle between each candidate extension direction and the original direction of the forklift, denoted as the candidate angle, where the original direction of the forklift is the driving direction of the target forklift before entering the candidate reversing point; sorting the candidate extension directions in ascending order of the angle difference between each candidate angle and a preset angle; and determining the candidate extension direction that ranks first and meets the preset conditions as the target extension direction; wherein the preset conditions include at least one of collision verification, the extendable length meeting the target extension distance, and the road width meeting a preset width threshold.

2. The method according to claim 1, characterized in that, The driving direction attribute includes forward, reverse, and bidirectional attributes; The positive attribute is used to limit the direction of the front of the target forklift to a first direction, the negative attribute is used to limit the direction of the front of the target forklift to a second direction, and the bidirectional attribute is used to allow the front of the target forklift to face either the first direction or the second direction. Wherein, the first direction points from the road start node to the road end node, the second direction points from the road end node to the road start node, the driving direction attribute does not restrict the target forklift from entering the road from either end of the road, and the driving direction attribute is used to determine the forward or backward mode of the target forklift on the road by constraining the direction of the forklift's front end.

3. The method according to claim 1 or 2, characterized in that, The step of traversing each adjacent node pair of the ordered node sequence and determining the target driving mode of the target forklift on each target road segment according to the starting point priority and direction priority inheritance strategy includes: For the first road segment, if the driving direction attribute allows the target forklift to pass in forward mode, then the corresponding target driving mode is determined to be forward mode; otherwise, the corresponding target driving mode is determined to be reverse mode. For any target road segment other than the first road segment, when the driving direction attribute of the target road segment is a single vehicle heading constraint: If the vehicle heading constraint conflicts with the driving mode inherited by the target forklift, then the target driving mode of the target forklift will be determined as the opposite driving mode to the target driving mode corresponding to the previous target road segment. If the vehicle heading constraint does not conflict with the driving mode inherited by the target forklift, then the target driving mode of the target forklift is determined to be the same as the target driving mode corresponding to the previous target road segment. When the driving direction attribute of the target road segment is bidirectional, the target forklift directly inherits the same driving mode as on the previous target road segment.

4. The method according to claim 1, characterized in that, The step of determining the target reversal point by extending the target extension distance along a target extension direction at the candidate reversal point includes: The target extension distance is determined by the following formula; ； in, Extend the distance to the target. The speed at which the target forklift arrives at the candidate reversal point. The preset maximum permissible deceleration is determined based on at least one of the following: cargo weight, cargo center of gravity height, and cargo friction coefficient. This is a preset distance constant.

5. The method according to claim 1, characterized in that, The method further includes: Obtain no-entry indication information for infeasible areas marked in the aforementioned road network topology; Based on the prohibition instruction information, on the target movement path, it is detected whether the vehicle outline of the target forklift interferes with the infeasible area; If the vehicle outline does not interfere with the infeasible area, then the target motion path is confirmed to be usable; If the vehicle outline interferes with the infeasible area, and the interference is located in the extension segment from the candidate reversal point to the target reversal point, then the candidate extension directions are switched sequentially according to the order of the candidate extension directions, and the target reversal point is re-determined. If all candidate extension directions cause interference, then the road direction or node direction combination that causes interference is masked, and the global path planning algorithm is re-executed.

6. The method according to claim 5, characterized in that, The step of detecting whether the vehicle outline of the target forklift interferes with the infeasible area on the target movement path according to the prohibition indication information includes: Along the complete path of the target motion path, multiple pose points are determined with a fixed step size; At each pose point, the heading and vehicle profile of the target forklift are determined based on the road direction of the target road segment where the pose point is located, the target driving mode corresponding to the target road segment, and the geometric dimensions of the target forklift. Check whether the vehicle profile overlaps with the infeasible region at any of the aforementioned pose points; The complete path includes the target road segment corresponding to the ordered node sequence and the extension segment between the candidate reversal point and the target reversal point.

7. The method according to claim 1, characterized in that, The preset angle is 90°, or the preset angle is within the range of 60° to 120°.

8. A path planning device for an unmanned forklift, characterized in that, The device includes: The road network information acquisition unit is used to acquire a topological road network composed of nodes and roads, wherein each road segment has a driving direction attribute, and the driving direction attribute is used to constrain the direction of the target forklift's front end. The path node planning unit is used to execute a global path planning algorithm on the topological road network according to the target start point and the target end point, and generate an ordered node sequence, wherein adjacent nodes in the ordered node sequence are connected by a target road segment; The driving mode determination unit is used to traverse each adjacent node pair of the ordered node sequence and determine the target driving mode of the target forklift on each target road segment according to the starting point priority forward and direction priority inheritance strategy; wherein, the target driving mode includes forward mode and reverse mode; in forward mode, the movement direction of the target forklift is the same as the front of the vehicle; in reverse mode, the movement direction of the target forklift is opposite to the front of the vehicle. The reversal point identification unit is used to identify the common nodes of the two adjacent target road segments as reversal candidate points if the target driving modes corresponding to the two adjacent target road segments are different. A reversing point extension unit is used to extend a target extension distance along a target extension direction at the reversing candidate point to determine a target reversing point. The target extension direction is different from the road direction of the two target road segments connecting the reversing candidate point. The target reversing point is used for the target forklift to move to the target reversing point after arriving at the reversing candidate point along the previous target road segment, and then switch between forward and reverse modes. According to the switched driving mode, it drives through the reversing candidate point into the next target road segment. The motion path generation unit is used to determine the target motion path of the target forklift based on the ordered node sequence, the target road segment, the target driving mode, the candidate turning points, and the target turning points. The step of determining the target reversing point by extending the target extension distance along a target extension direction at the candidate reversing point includes: determining all adjacent road directions of the candidate reversing point according to the topological road network; excluding the road directions of the two target road segments connected to the candidate reversing point from the adjacent road directions, and taking the remaining road directions as candidate extension directions; determining the angle between each candidate extension direction and the original direction of the forklift, denoted as the candidate angle, where the original direction of the forklift is the driving direction of the target forklift before entering the candidate reversing point; sorting the candidate extension directions in ascending order of the angle difference between each candidate angle and a preset angle; and determining the candidate extension direction that ranks first and meets the preset conditions as the target extension direction; wherein the preset conditions include at least one of collision verification, the extendable length meeting the target extension distance, and the road width meeting a preset width threshold.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory being used to store a computer program that, when executed by the processor, implements the method as described in any one of claims 1 to 7.

Images