Conveying system

Abstract

To provide a transportation system capable of suppressing an increase in the travel time of an automated guided vehicle, and inhibiting deterioration of work efficiency.SOLUTION: Plural observation devices observe a work area where an automated guided vehicle moves. Based on pieces of information obtained by the plural observation devices, an obstacle position information generation unit generates obstacle position information representing a position of an obstacle within the work area. The automated guided vehicle has a feature of acquiring the obstacle position information from the obstacle position information generation unit, a feature of obtaining a travel route, along which the automated guided vehicle moves from a current place to a target point while avoiding an obstacle, a feature of moving to the target point along the obtained travel route, and a feature of acquiring new obstacle position information from the obstacle position information generation unit even during a period until the automated guided vehicle reaches the destination, and correcting the current travel route on the basis of the new obstacle position information.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The present invention relates to a conveyance syste Mu m.

Background Art

[0002] An automated guided vehicle that travels unmanned within a work area travels without contacting obstacles while detecting surrounding obstacles using on-board sensors such as a camera and LiDAR. When an obstacle such as a pillar, another automated guided vehicle, a truck, a person or the like is detected in the traveling direction, the automated guided vehicle decelerates or stops. Patent Document 1 below discloses a control system for an automated guided vehicle using a SLAM (Simultaneous Localization and Mapping) guidance method.

Prior Art Literature

Patent Literature

[0003]

Patent Document 1

Summary of the Invention

Problem to be Solved by the Invention

[0004] An automated guided vehicle determines an optimal travel route to travel from the current location to a target location while avoiding obstacles, sets the optimal travel route in advance, and travels to the target location along the travel route. In some cases, a person, a forklift or the like may enter the preset travel route. When the automated guided vehicle detects an obstacle on the set travel route, it decelerates or stops. When the obstacle is removed from the set travel route, the automated guided vehicle resumes traveling. For this reason, the travel time to the target location becomes longer, leading to a decrease in work efficiency.

[0005] An object of the present invention is to provide a conveyance syste Mu m that can suppress an increase in the travel time of an automated guided vehicle and suppress a decrease in work efficiency. [Means for solving the problem]

[0006] According to one aspect of the present invention, Automated guided vehicles that move within the work area, A plurality of observation devices for observing within the work area, including at least one observation device fixedly positioned in the work area and a portable observation device, An obstacle location information generation unit generates obstacle location information indicating the location of obstacles within the work area based on information obtained from the aforementioned multiple observation devices. Equipped with, The aforementioned automated guided vehicle is A function to acquire the obstacle location information from the obstacle location information generation unit, Based on the acquired obstacle location information, the function determines a driving path that avoids obstacles from the current location to the target location. A function to move to the target location along the requested route, During the period until the target point is reached, the system also acquires new obstacle location information from the obstacle location information generation unit and modifies the current travel path based on the new obstacle location information. Equipped with 、 The obstacle location information generation unit creates multiple cost maps based on the type and location of the obstacle as the obstacle location information. The automated guided vehicle, as its travel route, determines the minimum cost route from its current location to the target location based on the cost map created by the obstacle location information generation unit. The obstacle location information generation unit divides the work area into a plurality of zones distributed in two dimensions, and when creating the plurality of cost maps, it assigns cost values ​​to each of the plurality of zones according to the presence or absence of obstacles and the type of obstacles, and creates the plurality of cost maps. The multiple cost maps are created such that the cost values ​​of at least some of the zones differ among the multiple cost maps. The automated guided vehicle determines its travel route using one cost map selected from the multiple cost maps. A transport system is provided. [Effects of the Invention]

[0008] Because automated guided vehicles (AGVs) correct their travel path based on new obstacle location information during their journey, they can suppress the length of travel time and improve work efficiency. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is a schematic plan view showing an overview of a transport system according to one embodiment. [Figure 2] FIG. 2 is a flowchart showing a procedure of processing performed by an obstacle position information generating unit. [Figure 3] FIGS. 3A to 3E are schematic diagrams for explaining processing executed by the obstacle position information generating unit. [Figure 4] FIG. 4 is a block diagram for explaining functions of an automated guided vehicle. [Figure 5] FIG. 5 is a flowchart showing a procedure of processing executed by an automated guided vehicle. [Figure 6] FIGS. 6A and 6B are plan views schematically showing an example of the position of an obstacle in a work area and the travel route of an automated guided vehicle. [Figure 7] FIG. 7 is a schematic diagram showing obstacle position information represented by a cost map. [Figure 8] FIGS. 8A to 8C are schematic diagrams for explaining a method of generating obstacle position information according to a comparative example. [Figure 9] FIGS. 9A and 9B are schematic diagrams showing zone-based cost maps generated as obstacle position information in a transport system according to another embodiment. [Figure 10] FIGS. 10A and 10B are schematic diagrams for explaining a method of detecting a person as an obstacle by a transport system according to still another embodiment. [Figure 11] FIG. 11A is a schematic diagram for explaining a method of detecting a person as an obstacle by a transport system according to still another embodiment, and FIG. 11B is a graph showing an example of a temporal change in the certainty of a detected object. [Figure 12] FIG. 12 is a block diagram showing functions of an automated guided vehicle according to still another embodiment. [Figure 13] FIG. 13 is a plan view schematically illustrating the automated guided vehicle according to the embodiment shown in FIG. 12. [Figure 14] FIGS. 14A and 14B are schematic diagrams respectively showing the relationship between the zone-based cost map of the transport system and automated guided vehicle according to the comparative example, the transport system and automated guided vehicle according to the embodiment shown in FIG. 12, and the automated guided vehicle. [Modes for carrying out the invention]

[0010] A transport system and automated guided vehicle according to one embodiment will be described with reference to Figures 1 to 7. Figure 1 is a schematic plan view showing an overview of a transport system according to one embodiment. The transport system according to this embodiment includes a plurality of observation devices 21, an automated guided vehicle (AGV) 30, and a higher-level control device 50. The AGV 30 determines its travel path based on obstacle location information acquired from the higher-level control device 50 and autonomously travels within the work area 20. For example, the work area 20 is a predetermined area in which the AGV 30 can move and transport goods to any location within the area based on commands.

[0011] Each of the multiple observation devices 21 observes an observable area within the work area 20. The observation results from the multiple observation devices 21 are input to the higher-level control device 50. The multiple observation devices 21 are, for example, cameras that capture images of the work area 20, and the obtained image data is input to the higher-level control device 50. Alternatively, other observation devices capable of acquiring information that can detect the type and location of obstacles within the work area 20 may be used as the multiple observation devices 21. For example, LiDAR (light detection and ranging) may be used as the observation device 21. The multiple observation devices 21 are arranged so as not to create blind spots within the work area 20. The area obtained by integrating the areas observed by each of the multiple observation devices 21 can also be called the work area 20.

[0012] The higher-level control device 50 includes an obstacle location information generation unit 51 and an input / output device 52. A manager or worker can give various commands to the higher-level control device 50 by operating the input / output device 52. The obstacle location information generation unit 51 detects the type and location of obstacles in the work area 20 based on information observed by multiple observation devices 21, such as image information, and generates obstacle location information.

[0013] Obstacles within the work area 20 include, for example, a person 25A, a forklift 25B operated by an operator, shelves 25C, and a truck 25D. Person 25A and the forklift 25B move within the work area 20. Shelves 25C are fixed within the work area 20. Truck 25D enters the work area 20 to load and unload cargo and stops within the work area 20. Truck 25D does not move during the loading and unloading operation.

[0014] The automated guided vehicle 30 is, for example, an automated guided vehicle (Automatic Guided Vehicle) that moves goods, or an automated guided forklift that holds and transports goods with its forks. The automated guided vehicle 30 uses short-range wireless communication such as Wi-Fi to send and receive various information with the higher-level control device 50.

[0015] Next, the processes performed by the obstacle location information generation unit 51 will be explained with reference to Figures 2 and 3A to 3E. Figure 2 is a flowchart showing the procedure of the processes performed by the obstacle location information generation unit 51 (Figure 1). Figures 3A to 3E are schematic diagrams illustrating the processes performed by the obstacle location information generation unit 51.

[0016] First, image data is acquired from each of the multiple observation devices 21 (step SA1). Figures 3A and 3B show the images defined by the image data acquired from each of the two observation devices 21. For example, the boundary between the floor and wall of the work area 20 (Figure 1), the boundary between two adjacent walls, and person 25A are captured within the image field of observation device 21. The observation ranges of the two observation devices 21 partially overlap. Therefore, one person 25A is captured in the two images shown in Figures 3A and 3B.

[0017] The obstacle location information generation unit 51 identifies the type and location of obstacles by separately analyzing image data acquired from each of the multiple observation devices 21 (step SA2). For example, a two-dimensional Cartesian coordinate system is defined within the work area 20. The coordinates of the boundary between the floor and walls of the work area 20 that are captured within the image boundary (Figures 3A and 3B) are known, and since the observation device 21 is fixed relative to the work area 20, the coordinates within the work area 20 can be calculated from the position within the image boundary.

[0018] For example, if a person is included in the image, the coordinates of the person's location can be calculated based on the position of their feet. For instance, in the image shown in Figure 3A, person 25A is detected at x=3500, y=4000, as shown in Figure 3C. Similarly, in the image shown in Figure 3B, person 25A is detected at x=3500, y=4000, as shown in Figure 3D. Even if there are obstacles other than people within the image, the coordinates of these obstacles can be calculated. The information defining the coordinates of obstacles is referred to as coordinate information.

[0019] Based on the image data acquired by each of the multiple observation devices 21, coordinate information is combined to generate obstacle position information covering the entire work area 20 (step SA3). Figure 3E shows an example of the combined coordinate information. In the combined coordinate information, the coordinates of person 25A, as defined in the coordinate information of Figures 3C and 3D, are combined into one. From the combined coordinate information, it can be seen that there is one person 25A at the position x=3500, y=4000. The obstacle position information generation unit 51 creates obstacle position information based on the combined coordinate information.

[0020] In this embodiment, a zone-based cost map is used as obstacle location information. The zone-based cost map divides the work area 20 into multiple zones and assigns a cost value to each of these zones. Zones where obstacles exist are assigned relatively higher cost values. Details of the zone-based cost map will be explained later with reference to Figure 7.

[0021] Next, the functions of the automated guided vehicle 30 will be described with reference to Figure 4. Figure 4 is a block diagram illustrating the functions of the automated guided vehicle 30. The automated guided vehicle 30 includes an obstacle location information acquisition unit 31, a zone-specific cost map creation and update unit 32, a travel path search unit 33, a control unit 34, and a travel mechanism 35.

[0022] The obstacle location information acquisition unit 31 acquires obstacle location information from the obstacle location information generation unit 51. The zone-specific cost map creation and update unit 32 creates or updates a zone-specific cost map based on the obstacle location information acquired by the obstacle location information acquisition unit 31. For example, it assigns cost values ​​to each of the multiple zones in the zone-specific cost map, or updates the cost values ​​that have already been assigned. The travel path search unit 33 calculates the travel path of the automated guided vehicle 30 from its current location to the target location based on the obstacle location information acquired by the obstacle location information acquisition unit 31. For example, it finds the minimum cost path (sometimes referred to as the "minimum cost travel path") based on the zone-specific cost map created or updated by the zone-specific cost map creation and update unit 32. The target location is commanded by the higher-level control device 50.

[0023] The control unit 34 controls the travel mechanism 35 to move along the travel path with the lowest cost determined by the travel path search unit 33. The travel mechanism 35 includes a motor and wheels, and the automated guided vehicle 30 moves when the control unit 34 drives the motor.

[0024] Next, the processes performed by the automated guided vehicle 30 will be explained with reference to Figure 5. Figure 5 is a flowchart showing the steps of the processes performed by the automated guided vehicle 30. First, the obstacle location information acquisition unit 31 (Figure 4) acquires obstacle location information from the obstacle location information generation unit 51 (Figure 4) (step SB1). Based on the acquired obstacle location information, the zone-specific cost map creation and update unit 32 creates or updates a zone-specific cost map (step SB2). Subsequently, the travel path search unit 33 (Figure 4) determines the travel path with the minimum cost from the current location to the target location based on the zone-specific cost map (step SB3). Once the travel path is determined, the control unit 34 (Figure 4) controls the travel mechanism 35 (Figure 4) to start traveling along the travel path (step SB4).

[0025] During the journey of the automated guided vehicle 30 to the target location, the AGV 30 periodically repeats the procedures from steps SB1 to SB4 (step SB5). While traveling along the minimum cost travel path determined in step SB3, the minimum cost travel path is periodically recalculated based on new obstacle location information, for example, every second (step SB3). If the recalculated travel path differs from the current travel path, the newly determined travel path is set as the minimum cost travel path, and the AGV travels along the new travel path (step SB4).

[0026] Next, the movement of the automated guided vehicle (AGV) 30 will be described with reference to Figures 6A and 6B. Figures 6A and 6B are schematic plan views showing an example of the location of obstacles in the work area 20 and the travel path of the AGV 30. Figure 6A shows the location of obstacles in the work area 20 at the start of travel. Two people 25A, a forklift 25B, a truck 25D, and a shelf 25C are present in the work area 20. Based on the current location of the obstacles, the AGV 30 determines the minimum cost travel path 22A from its current location to the target location 27.

[0027] As shown in Figure 6B, while the automated guided vehicle 30 is traveling, a person 25A moves and enters the travel path 22A. The obstacle location information generation unit 51 (Figure 1) generates the current obstacle location information (step SA3), and the automated guided vehicle 30 acquires new obstacle location information (step SB1). Based on the newly acquired obstacle location information, the automated guided vehicle 30 recalculates the travel path with the lowest cost to reach the target point 27 (step SB3). Since person 25A has entered the original travel path 22A, a new travel path 22B that avoids person 25A is set.

[0028] Next, the obstacle location information will be explained with reference to Figure 7. Figure 7 is a schematic diagram showing the obstacle location information represented in a cost map. The work area 20 is divided into multiple zones 40 arranged in a matrix. Each of the multiple zones 40 is, for example, a square with sides of about 1 to 2 m. Each of the multiple zones 40 is classified into one of four risk types: "high risk," "medium risk," "low risk," and "no risk." In the cost map shown in Figure 7, the level of risk is represented by shades of gray. Zones 40 with "high risk" are represented by the darkest shade, zones 40 with "low risk" are represented by the lightest shade, and zones 40 with "medium risk" are represented by an intermediate shade. Zones 40 with "no risk" are represented in white. In addition, the outlines of the forklift 25B, shelves 25C, and truck 25D are represented by white lines.

[0029] Based on the coordinate information of the obstacles shown in Figure 3E, the multiple zones 40 are classified into zones 40 with obstacles and zones 40 without obstacles. Zones 40 with obstacles are classified as "high risk".

[0030] Zones 40 located within a certain distance from the zone 40 where person 25A was detected, for example, multiple zones 40 adjacent to the zone 40 where person 25A was detected in the row, column, and diagonal directions, are classified as having a "medium risk" risk level. The multiple "medium risk" zones 40 are arranged to surround the zone 40 where person 25A was detected in a ring. The risk level of the multiple zones 40 adjacent to the outer side of the ring-shaped "medium risk" zones 40 is classified as having a "low risk" risk level. In other words, the zone 40 where person 25A was detected is surrounded by the "medium risk" zones 40, and the "low risk" zones 40 surround the outside of that.

[0031] For zones 40 where non-human obstacles, such as forklifts 25B, shelves 25C, or trucks 25D, are detected, no "medium risk" zone is assigned. For consecutive zones 40 where each obstacle is detected, the zones 40 adjacent to the outermost zone 40 in the row, column, and diagonal directions are classified as "low risk." In other words, consecutive zones 40 where each non-human obstacle is detected are surrounded by multiple "low risk" zones 40.

[0032] Multiple zones 40 are assigned cost values ​​according to their risk level. For example, a cost value of "30,000" is assigned to a "high-risk" zone 40, a cost value of "10" to a "medium-risk" zone 40, a cost value of "5" to a "low-risk" zone 40, and a cost value of "1" to a "no-risk" zone 40. A cost map divided into multiple zones 40, with a cost value assigned to each zone 40, is called a "zone-specific cost map." In step SB3 (Figure 5), the minimum cost path of the zone-specific cost map shown in Figure 7 is calculated. The minimum cost path can be found, for example, using Dijkstra's algorithm.

[0033] Next, we will explain the excellent effects of the embodiments shown in Figures 1 to 7. In the above embodiment, while the automated guided vehicle 30 (Figure 1) is traveling to reach the target point, it periodically acquires new obstacle location information, for example, every second (step SB1), and revises the minimum cost travel path (step SB3). As shown in Figure 6B, if an obstacle (person 25A) enters the current minimum cost travel path 22A, the current travel path 22A can be modified without approaching the obstacle to a detectable distance, and a new travel path 22B that avoids the obstacle can be determined. For example, if an obstacle blocks the current minimum cost travel path near the target point, the minimum cost travel path can be modified before reaching the location of the obstacle. This avoids wasted time such as deceleration and stopping of the automated guided vehicle 30, and improves work efficiency.

[0034] Furthermore, in the above embodiment, since multiple observation devices 21 (Figure 1) are installed in the work area 20, a wider area within the work area 20 can be observed compared to the case where observation devices are mounted only on the automated guided vehicle 30. The type and location of obstacles in areas that cannot be observed by observation devices mounted on the automated guided vehicle 30 can be detected. Considering the type and location of these obstacles, the travel route with the lowest cost can be determined.

[0035] In the above embodiment, as shown in Figure 7, a zone-specific cost map divided into multiple zones 40 is used as obstacle location information. Therefore, the amount of data for obstacle location information is reduced compared to the case where a cost map defined by a small point cloud is used. As a result, the computational load required to search for the minimum cost travel route can be reduced. This allows for the rapid search for the minimum cost travel route even while traveling.

[0036] In the above embodiment, as shown in Figure 7, the zone 40 where obstacles other than person 25A are detected is surrounded by a single layer of "low risk" zone 40. In contrast, the zone 40 where person 25A is detected is surrounded by a double layer of "medium risk" zone 40 and "low risk" zone 40. Therefore, the minimum cost travel path is determined so as not to travel near person 25A. This further enhances the safety of person 25A.

[0037] Next, we will describe further advantages of the above embodiment in comparison with the comparative examples shown in Figures 8A to 8C. Figures 8A to 8C are schematic diagrams illustrating the method for generating obstacle position information using the comparative examples. Images shown in Figures 8A and 8B are acquired by two observation devices 21. The same person 25A is captured in both images. These two images are combined to generate the single image shown in Figure 8C. By analyzing the combined image, the type and position (coordinates) of obstacles such as person 25A are identified.

[0038] In contrast, in the above embodiment, the individual images shown in Figures 3A and 3B are analyzed to generate the coordinate information shown in Figures 3C and 3D, and the two sets of coordinate information are combined to identify the location (coordinates) of the obstacle. The computational load of the process of combining multiple sets of coordinate information is smaller than the computational load of the process of combining multiple images. Therefore, the computational load for generating obstacle location information is reduced, and obstacle location information can be generated in a shorter time.

[0039] Next, a modified example of the embodiment described above will be explained with reference to Figures 1 to 7. In the embodiments shown in Figures 1 to 7, the automated guided vehicle (AGV) 30 (Figure 1) acquires obstacle location information from the higher-level control device 50 (Figure 1) (step SB1) and determines the minimum cost travel route to the target location (step SB3). As a variation, the higher-level control device 50 may obtain the current position of the AGV 30 and determine the minimum cost travel route from the current position to the target location. In this case, the higher-level control device 50 may transmit information specifying the minimum cost travel route to the AGV 30.

[0040] In the above embodiment, a zone-based cost map (Figure 7) is used as obstacle location information to calculate the travel route with the lowest cost. Alternatively, obstacle location information other than the zone-based cost map may be used to search for a travel route that avoids obstacles. When searching for a travel route, the type of obstacle and the width of the drivable area on the route may be taken into consideration when determining the travel route. Even in this case, by periodically acquiring obstacle location information not only around the vehicle but also throughout the entire work area 20 (Figure 1), it is possible to periodically correct the travel route while the automated guided vehicle 30 is reaching the target point.

[0041] In the above embodiment, the minimum cost travel route is reviewed periodically until the automated guided vehicle 30 reaches the target location (step SB3). Alternatively, the minimum cost travel route may be reviewed not periodically, but as needed. For example, when a change occurs in the cost value of at least one of the zones in the zone-based cost map, the higher-level control device 50 may send a new zone-based cost map to the automated guided vehicle 30. When the automated guided vehicle 30 receives the new zone-based cost map, it may perform a process to review the minimum cost travel route. Alternatively, the minimum cost travel route may be reviewed every time a predetermined distance is traveled. In this way, the minimum cost travel route should be reviewed at least once, preferably multiple times, while the automated guided vehicle 30 is traveling toward the target location.

[0042] In the above embodiment, obstacle location information, i.e., a zone-specific cost map, is created or updated based on the observation results of multiple observation devices 21 fixedly positioned within the work area 20. However, other observation devices may also be used in combination. For example, obstacle location information may be created by using observation results from portable observation devices, observation devices mounted on the vehicle itself, and observation devices mounted on other vehicles moving within the work area 20.

[0043] Next, a transport system and automated guided vehicle according to another embodiment will be described with reference to Figures 9A and 9B. Hereafter, the description of components common to the transport system and automated guided vehicle described with reference to Figures 1 to 7 will be omitted.

[0044] Figures 9A and 9B are schematic diagrams showing zone-specific cost maps generated as obstacle location information in the transport system according to this embodiment. The obstacle location information generation unit 51 (Figure 1) generates multiple, for example, two zone-specific cost maps (Figures 9A and 9B) based on the same image acquired by multiple observation devices 21. The distribution of risk types in the two zone-specific cost maps is the same. However, the cost values ​​assigned to each of the multiple zones 40 are different.

[0045] For example, in the zone-based cost map shown in Figure 9A, a risk value of "30,000" is assigned to zone 40 as "high risk," a risk value of "10" is assigned to zone 40 as "medium risk," a risk value of "5" is assigned to zone 40 as "low risk," and a risk value of "1" is assigned to zone 40 as "no risk." The zone-based cost map shown in Figure 9A is called the zone-based cost map for "normal mode." In the other zone-based cost map shown in Figure 9B, the cost value assigned to zone 40 as "medium risk" is reduced from "10" to "7," and the risk value assigned to zone 40 as "low risk" is reduced from "5" to "1," making it the same as the cost value of "1" for zone 40 as "no risk." The zone-based cost map shown in Figure 9B is called the zone-based cost map for "cycle time priority mode."

[0046] The obstacle location information generation unit 51 (Figure 1) transmits one of two zone-specific cost maps, "normal mode" and "cycle time priority mode," to the automated guided vehicle 30. The administrator or operator specifies which zone-specific cost map to transmit by operating the input / output device 52 (Figure 1).

[0047] In the "Normal Mode" zone-based cost map, the cost value of zone 40 with "no risk" is smaller than the cost values ​​of the other zones 40. Therefore, the minimum cost route 22A is determined to preferentially pass through zone 40 with "no risk". In contrast, in the "Cycle Time Priority Mode" zone-based cost map, zones 40 with "no risk" and zones 40 with "low risk" are assigned the same cost value. Therefore, zones 40 with "low risk" are not distinguished from zones 40 with "no risk," and the minimum cost route 22A is determined accordingly. Consequently, using the "Cycle Time Priority Mode" zone-based cost map may result in a shorter route being determined as the shortest cost route compared to using the "Normal Mode" zone-based cost map.

[0048] For example, if it is guaranteed that there are no civilians in the work area 20, such as at night, it is preferable to send the zone-based cost map for "cycle time priority mode" to the automated guided vehicle (AGV) 30 instead of the zone-based cost map for "normal mode". The AGV 30 will then travel along a shorter route compared to when using the zone-based cost map for "normal mode".

[0049] Next, we will describe the excellent effects of the embodiments shown in Figures 9A and 9B. In this embodiment, if it is guaranteed that there are no civilians in the work area 20 and sufficient safety is ensured, the minimum cost travel route can be determined using the zone-specific cost map in "cycle time priority mode". As a result, the travel route of the automated guided vehicle 30 can be shortened, and work efficiency can be increased.

[0050] Next, with reference to Figures 10A and 10B, a conveying system according to another embodiment will be described. The following description will omit details of components common to the conveying systems and automated guided vehicles described with reference to Figures 1 to 7.

[0051] Figures 10A and 10B are schematic diagrams illustrating how the transport system according to this embodiment detects a person as an obstacle. As shown in Figure 10A, a person 25A is detected within one zone 40A of the zone-based cost map. The higher-level control device 50 has a function to detect when a person 25A crosses the boundary line of zone 40 and moves to the adjacent zone 40. As shown in Figure 10B, when a person 25A moves from zone 40A to the adjacent zone 40B, the higher-level control device 50 detects that a person 25A has moved from zone 40A to zone 40B. For example, an observation device 21 may be placed on the ceiling of the work area 20 (Figure 1) to monitor the movement of person 25A. The higher-level control device 50 determines that a person 25A is in the destination zone 40B.

[0052] Next, we will describe the excellent effects of the embodiments shown in Figures 10A and 10B. The process of detecting people in an image by analyzing the image acquired by the observation device 21 (Figure 1) can be performed using artificial intelligence (AI). However, the accuracy of AI-based person detection is not 100%, and there may be cases where a person is not detected even if they are present. For example, detection may be difficult depending on the angle at which the person is photographed and their posture.

[0053] In this embodiment, if a person 25A is detected in a zone 40A, and it is detected that person 25A has moved to another zone 40B, it is determined that person 25A is in zone 40B even if no person is detected in zone 40B. Therefore, even if the AI's person detection is unstable, the detection accuracy of person 25A can be improved.

[0054] Next, with reference to Figures 11A and 11B, a conveying system according to another embodiment will be described. The following description will omit details of components common to the conveying systems and automated guided vehicles described with reference to Figures 1 to 7.

[0055] Figure 11A is a schematic diagram illustrating how the transport system according to this embodiment detects people as obstacles. In one zone 40A of the zone-based cost map, 25 people are detected. When people are detected by image analysis, the probability that the detected object is a person (hereinafter referred to as "likelihood") is usually calculated. If the likelihood is above a certain threshold, the detected object is determined to be a person.

[0056] Figure 11B is a graph showing an example of the change in the likelihood of detecting an object over time. The horizontal axis represents time, and the vertical axis represents likelihood. In the graph in Figure 11B, the solid line shows the instantaneous value of likelihood, and the dashed line shows the time-moving average. When the likelihood is greater than or equal to the threshold Th, the detected object is determined to be a person. The likelihood fluctuates depending on the angle and posture of the person being photographed. For example, if a person is crouching or lying down, the likelihood may decrease and fall below the threshold Th. When the likelihood falls below the threshold Th, the detected object is determined not to be a person. In the example shown in Figure 11B, when detecting a person using the instantaneous value of likelihood, the object detected in periods T1 and T2 is determined not to be a person.

[0057] In this embodiment, a time-moving average of the probability is calculated, and a determination of whether the object is a person is made based on the time-moving average. For example, if the time-moving average of the probability is greater than or equal to a threshold Th, the detected object is determined to be a person.

[0058] Next, we will describe the excellent effects of the embodiment shown in Figures 11A and 11B. Even if a person's posture changes and the probability temporarily falls below the threshold Th, if the time spent below the threshold Th is short, the time-moving average of the probability remains above the threshold. Therefore, even if a person's posture changes and the probability temporarily decreases, it is still possible to detect them.

[0059] Next, with reference to Figures 12 to 14B, we will describe other embodiments of transport systems and automated guided vehicles. Hereafter, we will omit explanations of components common to the transport systems and automated guided vehicles described with reference to Figures 1 to 7.

[0060] Figure 12 is a block diagram showing the functions of the automated guided vehicle 30 according to this embodiment, and Figure 13 is a schematic plan view of the automated guided vehicle 30. In addition to the obstacle position information acquisition unit 31, travel path search unit 33, control unit 34, and travel mechanism 35 shown in Figure 4, the automated guided vehicle 30 is equipped with obstacle detection sensors 36 and a mode command receiving unit 38. The obstacle detection sensors 36 (Figure 13) are mounted, for example, at diagonal positions of the automated guided vehicle 30 in a plan view, and detect the distance to obstacles around the automated guided vehicle 30.

[0061] The mode command receiving unit 38 receives the driving mode of the automated guided vehicle 30 from the higher-level control device 50. The driving modes include "normal mode" and "cycle time priority mode," as in the embodiment described with reference to Figures 9A and 9B.

[0062] A virtual bumper area 37 (Figure 13) is set around the automated guided vehicle 30. The control unit 34 determines whether an obstacle detected by the obstacle detection sensor 36 is located inside the virtual bumper area 37. If an obstacle is detected inside the virtual bumper area 37, the control unit 34 controls the driving mechanism 35 to decelerate or stop the automated guided vehicle 30.

[0063] The control unit 34 changes the size of the virtual bumper area 37 depending on the driving mode. When the driving mode is cycle time priority mode, the virtual bumper area 37 is made smaller than when it is normal mode.

[0064] Next, the excellent effects of this embodiment will be described with reference to Figures 14A and 14B. Figures 14A and 14B are schematic diagrams showing the relationship between zone-based cost maps and automated guided vehicles 30. The work area 20 (Figure 1) is divided into multiple zones 40, and each zone 40 is classified into risk categories: "high risk," "low risk," and "no risk." The "medium risk" zone 40 shown in Figure 7 does not exist in Figures 14A and 14B. In Figures 14A and 14B, the "high risk" zone 40 is shown in the darkest color, the "medium risk" zone 40 is shown in a lighter color, and the "no risk" zone 40 is shown in white.

[0065] When the driving mode is set to "cycle time priority mode," as shown in Figure 9B, the same risk value of "1" is assigned to both the "low risk" zone 40 and the "no risk" zone 40. Therefore, the "low risk" zone 40 is also selected as the minimum cost driving route. Obstacles, such as forklifts 25B and trucks 25D, exist near the "low risk" zone 40. Therefore, the automated guided vehicle 30 will pass near these obstacles.

[0066] When passing near an obstacle, as shown in Figure 14A, if the obstacle is detected inside the virtual bumper area 37, the automated guided vehicle 30 will decelerate or stop. In this embodiment, as shown in Figure 14B, the virtual bumper area 37 is made smaller in "cycle time priority mode," making it less likely for obstacles to be detected inside the virtual bumper area 37, and allowing the automated guided vehicle 30 to pass near obstacles without decelerating.

[0067] When it is guaranteed that there are no civilians in the work area 20 and safety is ensured, the driving mode is set to "cycle time priority mode". This avoids unnecessary deceleration and stopping of the automated guided vehicle 30, thereby improving work efficiency.

[0068] The embodiments described above are illustrative, and it goes without saying that partial substitution or combination of the configurations shown in different embodiments is possible. Similar effects and benefits from similar configurations in multiple embodiments will not be mentioned sequentially for each embodiment. Furthermore, the present invention is not limited to the embodiments described above. For example, it will be obvious to those skilled in the art that various modifications, improvements, and combinations are possible. [Explanation of Symbols]

[0069] 20 Work Areas 21 Observation equipment Routes 22A and 22B 25A people 25B Forklift 25C shelf 25D Track 27 Target point 30 Automated Guided Vehicles 31 Obstacle location information acquisition unit 32. Zone-Specific Cost Map Creation and Update Section 33. Route Search Unit 34 Control Unit 35. Running mechanism 36 Obstacle detection sensor 37 Virtual Bumper Area 38 Mode command receiving unit Zones 40, 40A, and 40B 50 Higher-level control unit 51 Obstacle location information generation unit 52 Input / Output Devices

Claims

1. Automated guided vehicles that move within the work area, The observation devices include at least one fixed observation device and one portable observation device located in the work area, and a plurality of observation devices for observing within the work area. An obstacle location information generation unit generates obstacle location information indicating the location of obstacles within the work area based on information obtained from the aforementioned multiple observation devices. Equipped with, The aforementioned automated guided vehicle is A function to acquire the obstacle location information from the obstacle location information generation unit, Based on the acquired obstacle location information, the function determines a driving path that avoids obstacles from the current location to the target location. A function to move to the target location along the requested route, During the period until the target point is reached, the system also acquires new obstacle location information from the obstacle location information generation unit and modifies the current travel path based on the new obstacle location information. Equipped with, The obstacle location information generation unit, The aforementioned work area is divided into multiple zones distributed in two dimensions, As the aforementioned obstacle location information, cost values ​​are assigned to each of the multiple zones according to the presence or absence of obstacles and the type of obstacles to create multiple cost maps. The multiple cost maps are created such that the cost values ​​of at least some of the zones differ among the multiple cost maps. The aforementioned automated guided vehicle is a transport system that uses one cost map selected from the multiple cost maps to determine the minimum cost route from the current location to the target location, and determines the minimum cost route as the travel route.

2. The transport system according to claim 1, wherein the obstacle position information generation unit detects a person based on information obtained from the plurality of observation devices, and when it detects that the detected person has crossed the boundary line of the plurality of zones, it determines that there is a person in the zone after the movement.

3. The transport system according to claim 1, wherein the obstacle position information generation unit detects a person based on information obtained from the plurality of observation devices, and determines that it is a person if the time-moving average of the certainty of person detection is greater than or equal to a threshold.

4. A virtual bumper area is defined around the aforementioned automated guided vehicle. The aforementioned automated guided vehicle is The function slows down or stops when an obstacle is detected inside the virtual bumper area, A function to change the size of the virtual bumper area The transport system according to claim 1, further comprising:

Images