Apparatus and method for predicting collision of logistics robot

Abstract

Disclosed are an apparatus and method for predicting collision of a logistics robot, the apparatus comprising: a collection unit for collecting process state information, a learning unit for learning a collision prediction criterion for each collision type based on the process state information, a prediction unit for predicting the collision prediction criterion, and an output unit for outputting the collision prediction criterion; the prediction unit is used for predicting occurrence of collision of at least one of the plurality of logistics robots based on a collision prediction standard and the process state information collected after learning, and the output unit is used for outputting a control signal corresponding to a prediction result.

Description

TECHNICAL FIELD

[0001] The present invention relates to a collision prediction device and method for a logistics robot for predicting and preventing a collision of a logistics robot in a working boundary including a plurality of facilities and logistics robots. BACKGROUND

[0002] In a conventional logistics warehouse and factory and a smart factory which manufactures products having different specifications using various components, logistics robots are introduced to achieve flexible and efficient supply and transfer of components.

[0003] Logistics robots are a concept collectively referred to as Autonomous Mobile Robots (AMRs) and Automated Guided Vehicles (AGVs), and such logistics robots are capable of performing movement and work under the control of a control device.

[0004] In a smart factory, in order to perform a task assigned from a control device, a logistics robot can travel along an optimal movement path based on path planning.

[0005] A plurality of logistics robots can be provided in a smart factory, and when movement paths of the plurality of logistics robots overlap, a collision between the logistics robots can occur. For example, in a narrow passage of a smart factory, when one logistics robot moves downward and another logistics robot moves upward, the two logistics robots can be in a stalemate and collide. In addition, the logistics robots can collide with facilities provided in the smart factory during movement in the smart factory.

[0006] Therefore, when a collision of logistics robots occurs, a loss due to a malfunction of the logistics robots and a significant impact on the progress of the entire process can occur, resulting in a huge loss. Therefore, there is a need to propose a solution capable of predicting and avoiding a collision risk in advance before a collision of logistics robots.

[0007] The above description is intended only to assist in understanding the background of the present invention and is not intended to represent the related art falling within the scope of the knowledge of those skilled in the art. SUMMARY

[0008] TECHNICAL PROBLEM

[0009] The present invention is directed to providing a collision prediction device and method for a logistics robot which pre-learns a collision prediction criterion of a logistics robot through process state information in a working boundary and predicts and prevents a collision of a logistics robot according to a learning result.

[0010] The objects of the present application are not limited to the above-mentioned objects, and other objects not mentioned will be clearly understood by those skilled in the art from the following description.

[0011] Means for solving the problem

[0012] To achieve the above object, according to the present application, there is provided a collision prediction device for a logistics robot, the collision prediction device comprising a collection part, a learning part, a prediction part, and an output part, the collection part being configured to collect process state information, the process state information including facility information about a plurality of facilities disposed in a predetermined work boundary and logistics robot work information about a plurality of logistics robots, the plurality of logistics robots being disposed in the work boundary and moving through at least one or a part of a plurality of facility areas in which the plurality of facilities are disposed; the learning part being configured to learn a collision prediction criterion for each collision type based on the process state information, the collision type including a collision between any one of the plurality of facilities and any one of the plurality of logistics robots and a collision between the plurality of logistics robots; the prediction part being configured to predict occurrence of a collision of at least one of the plurality of logistics robots based on the process state information collected after learning and the learned collision prediction criterion; and the output part being configured to output a control signal corresponding to the prediction result.

[0013] For example, the learning part can be configured to label the process state information according to a time period, and learn the collision prediction criterion for each collision type based on the labeled process state information.

[0014] For example, the learning part can be configured to classify the labeled process state information according to whether a predetermined distance condition of a target object is satisfied for the plurality of logistics robots, the target object including the facility and another of the plurality of logistics robots.

[0015] For example, the learning part can be configured to learn the collision prediction criterion based on process state information corresponding to a point in time before a predetermined time interval from a point in time at which the predetermined distance condition is satisfied in the labeled process state information.

[0016] For example, the learning part can be configured to perform learning when any one of the plurality of logistics robots satisfies the predetermined distance condition.

[0017] For example, the learning part can be configured to learn the collision prediction criterion based on process state information corresponding to a point in time before a predetermined time interval from a point in time at which the predetermined distance condition is not satisfied in the labeled process state information.

[0018] For example, the facility information can include at least one of coordinates of a facility area corresponding to the facility in the work boundary or a task state of the facility.

[0019] For example, the facility information can vary according to a working characteristic of the facility.

[0020] For example, the logistics robot working information can include at least one selected from the group of: a speed, a moving direction, a moving path, a size, a normal working state, a sensor detection result, a braking specification, a real-time clock, a characteristic of an article loaded by each of the plurality of logistics robots, an inter-robot distance between some of the plurality of logistics robots, and a working priority order among the plurality of logistics robots.

[0021] For example, the working procedure state information can further include detection information measured by a detection device for detecting the facility and the logistics robot in the working boundary.

[0022] For example, the control signal can include a control signal for expressing an alert notification based on the prediction result.

[0023] For example, when the collision type corresponds to a collision between the plurality of logistics robots, the control signal can include a signal for controlling at least one of a speed or a moving direction of at least one of the plurality of logistics robots.

[0024] For example, the control signal can include a signal for controlling the plurality of logistics robots such that the plurality of logistics robots work according to the working priority order among the plurality of logistics robots.

[0025] For example, when the collision type corresponds to a collision between any one of the plurality of facilities and any one of the plurality of logistics robots, the control signal can include a signal for transmitting to the facility and controlling a task state of the facility.

[0026] For example, when the collision type corresponds to a collision between any one of the plurality of facilities and any one of the plurality of logistics robots, the control signal can include a signal for transmitting to the logistics robot and controlling at least one of a speed or a moving direction of the logistics robot.

[0027] To achieve the above object, according to the present application, there is provided a collision prediction method for a logistics robot, the collision prediction method including: collecting process state information, the process state information including facility information about a plurality of facilities disposed in a preset work boundary and logistics robot work information about a plurality of logistics robots disposed in the work boundary and moving through at least one or a part of a plurality of facility areas in which the plurality of facilities are disposed; learning a collision prediction criterion for each collision type based on the process state information, the collision type including a collision between any one of the plurality of facilities and any one of the plurality of logistics robots and a collision between the plurality of logistics robots; predicting occurrence of a collision of at least one of the plurality of logistics robots based on the process state information collected after the learning and the learned collision prediction criterion; and outputting a control signal corresponding to the prediction result.

[0028] Effects of Invention

[0029] According to the above-described embodiments of the present application, a collision of a logistics robot can be predicted in advance. Specifically, by learning a collision prediction criterion for a collision type that can occur in a work boundary, the performance of collision prediction can be improved.

[0030] Further, a logistics robot or a facility is controlled according to a result of collision prediction, thereby preventing a collision from occurring.

[0031] Through the above effects, a logistics robot or a facility can be prevented from being damaged, and further, a process can be continued in a stable manner by preventing a process delay due to occurrence of a collision event.

[0032] Effects that can be obtained from the present application are not limited to the above-described effects. Further, other effects not described herein will become apparent to those skilled in the art from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a block diagram illustrating a configuration example of a work boundary suitable for an embodiment of the present application.

[0034] Figure 2 is a schematic diagram illustrating an implementation example of a work boundary suitable for an embodiment of the present application.

[0035] Figure 3 is a block diagram illustrating a configuration example of a control device suitable for an embodiment of the present application.

[0036] Figure 4 is a block diagram illustrating a configuration example of a logistics robot suitable for an embodiment of the present application.

[0037] Figure 5FIG. 7 is a sequence diagram illustrating a process of predicting a collision of a logistics robot according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0038] Hereinafter, embodiments described in the present specification will be described in detail with reference to the accompanying drawings. In all the drawings, the same or similar elements are designated with the same reference numerals, and redundant descriptions of the same or similar elements will be omitted. The suffix "module" or "part" used in the following description is given or used only in order to facilitate explanation of the present specification, and has no distinct meaning or role in itself. Furthermore, in describing the embodiments disclosed in the present specification, if it is determined that the detailed description of the known technical to related to the present disclosure obscures the subject matter of the embodiments disclosed in the present specification, a detailed description of the known technical will be omitted. Furthermore, the accompanying drawings are merely provided to facilitate the understanding of the embodiments disclosed in the present specification, and do not limit the technical idea of the embodiments disclosed in the present specification. In addition, it should be understood that the present disclosure includes all modifications, equivalents, and substitutions included in the spirit and scope of the present disclosure.

[0039] Although the terms "first", "second", and the like used in the specification can be used to describe various elements, the elements are not interpreted as being limited by the terms. The terms are used only to distinguish one element from another.

[0040] It should be understood that when an element is referred to as being "coupled" or "connected" to another element, the element can be directly coupled or connected to the other element or intervening elements or intervening components can be present between the elements. In contrast, it should be understood that when an element is referred to as being "directly coupled" or "directly connected" to another element, then there are no intervening elements or intervening components present between the elements.

[0041] As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0042] In the present specification, it will be understood that terms such as "include", "have", and the like are intended to indicate the presence of features, numbers, steps, actions, elements, components, or combinations thereof disclosed in the specification, and are not intended to preclude the possibility of the presence or possibility of addition of one or more other features, numbers, steps, actions, elements, components, or combinations thereof.

[0043] Further, the units or control units included in the names of the internal configurations of the logistics robots or the control apparatus are only terms widely used to name controllers that control specific functions, and do not mean general functional units. For example, each controller can include a modem / transceiver that communicates with other controllers or sensors to control functions responsible for, a memory that stores an operating system or logic instructions and input / output information, and at least one processor that performs judgments, calculations, and decisions required to control functions responsible for. According to the implementation, a single processor can be responsible for the operation of multiple controllers.

[0044] First, referring to Figure 1 and Figure 2 a configuration of a work boundary in which the logistics robots are disposed and work according to an embodiment is described.

[0045] Figure 1 is a block diagram illustrating a configuration example of a work boundary applicable to the inventive embodiment.

[0046] Referring to Figure 1 , the work boundary 100 can include logistics robots 110, facilities 120, detection apparatuses 130, and a control apparatus 140.

[0047] According to the production process of the product and the target production speed, a plurality of logistics robots 110, a plurality of facilities 120, and a plurality of detection apparatuses 130 can be disposed in the work boundary 100. The work boundary 100 can be implemented as a smart factory, and the plurality of facilities can be implemented as production apparatuses, but are not necessarily limited thereto. Hereinafter, each element will be described.

[0048] First, the logistics robots 110 can include autonomous mobile robots (hereinafter, referred to as "AMRs" for convenience) and automated guided vehicles (hereinafter, referred to as "AGVs" for convenience). According to the work strategy of the logistics robots 110, only one type (one of AGVs and AMRs) can work in the work boundary 100, or both types (AGVs and AMRs) can work together in the work boundary 100.

[0049] The AGV generally performs a required work (such as moving, changing direction, and stopping) in the work boundary 100 by recognizing and following a guide facility (which is provided on the ground to provide a guide for the AGV). Herein, the guide facility can refer to an optically recognizable marker (such as a marker point, a two-dimensional code), a short-range non-contact recognition tag (such as an NFC tag, an RFID tag), a magnetic stripe, and a guide wire, but these are merely exemplary and are not necessarily limited thereto. The guide facility can be continuously provided on the ground, or can be discontinuously provided at intervals from each other. Since the AGV basically performs a work by recognizing and following the guide facility, it is necessary to install the guide facility in advance before the work. When the AGV needs to move to a new path or the original path needs to be modified, new construction or modification of the guide facility must be physically performed. In addition, the AGV does not deviate from the path provided by the guide facility. Therefore, when an obstacle is detected on or near the path, the AGV generally stops until the detected obstacle disappears or a separate control is received. In the work of the AGV, the control device 140 needs to control the AGV based on the guide facility, and can transmit an instruction such as "travel to the third marker recognized" or "change the travel direction by 90 degrees when the third marker is recognized" to the AGV based on a single instruction or a task instruction (for example, collection, supply, charging, and patrol) including a plurality of instructions.

[0050] The AMR is capable of determining a current position (i.e., localization) through environmental detection, and is capable of performing autonomous path setting (path planning) using the localization and a map, which is the most distinctive feature from the AGV. Accordingly, when the AMR shares a map having coordinate compatibility with the control device 140, the control device 140 can control the AMR by issuing a path instruction based on coordinates to the AMR. In addition, when an obstacle is detected during travel, the AMR can autonomously set an obstacle avoidance path to avoid the obstacle, and return to the original path. The function of the control device 140 to set an AMR path using one or more way coordinates can be referred to as global path planning. The function of the AMR to set a movement path or an obstacle avoidance path between the way coordinates according to the global path planning can be referred to as local path planning.

[0051] Next, for example, the facility 120 can refer to a device (such as a robot arm, a conveyor) that performs a product production process in the work boundary 100. In a broad sense, when the production process is performed by a person, the facility 120 can refer to a device provided to assist the execution of a task (for example, entry or exit of the logistics robot 110). The device provided to assist the execution of a task can be a device for detecting the state of a designated position at which the logistics robot 110 can drop or pick up a pallet in an area in which a specific production process is performed, a device for determining the progress of the process, or a device for blocking the entry and exit of the area, but is not necessarily limited thereto.

[0052] For example, the facility 120 can be controlled by a Programmable Logic Controller (PLC), and can communicate with the control device 14 in terms of progress of a process.

[0053] The detection device 130 can perform a function of acquiring information for determining a situation in the work boundary 100, and transmitting the information to the control device 140. For example, the detection device 130 can include a camera and a proximity sensor, but is not necessarily limited thereto.

[0054] The control device 140 can communicate with the above-described elements 110, 120, and 130 to acquire information required for the work of the work boundary 100 or to control each of the elements. For example, the control device 140 can perform scheduling, path setting, task assignment, process management for each product, and material management of the logistics robot 110.

[0055] In an embodiment, the control device 140 can include an AMR / AGV Control System (ACS) that controls a surrounding process facility based on a position of an AGV / AMR and controls task execution based on the AGV / AMR, and a Mobile Robot Integrated Monitoring System (MoRIMS) that controls two or more AMR / AGV control systems in an integrated manner. The Mobile Robot Integrated Monitoring System can perform a function of acquiring a state and a path of all logistics robots 110 in the work boundary 100 from each of a plurality of AMR / AGV control systems, performing logistics flow setting, and traffic control. For example, when the AMR / AGV control system (ACS) is provided with logistics robots of the same manufacturer or the same model, the Mobile Robot Integrated Monitoring System can perform integrated control for collision avoidance through traffic dispersion control between various types based on information obtained through a plurality of AMR / AGV control systems (ACS), such as analysis of a bottleneck level for an overlapping / crossing area, travel acceleration / deceleration control, and obstacle avoidance path regeneration.

[0056] Further, the Mobile Robot Integrated Monitoring System can have a Manufacturing Execution System (MES) as a higher-level controller, and the Manufacturing Execution System (MES) can be linked with an Advanced Planning & Scheduling (APS) system.

[0057] It is to be understood that, in addition to the elements 110, 120, 130, and 140 of the working boundary 100 described above, devices for inter-element communication, such as beacons, repeaters, and access points (APs), charging devices for charging the logistics robots 110, loading spaces for storing or loading components, spaces for storing final products or intermediate products, traffic lights, automatic doors, and standby spaces for idle logistics robots 110, can be appropriately provided in the working boundary 100.

[0058] Hereinafter, an example of a working boundary 100 suitable for an embodiment of the present application will be described with reference to Figure 2 An example of an implementation of the working boundary 100 suitable for an embodiment of the present application will be described.

[0059] Figure 2 is an example illustrating an implementation of a working boundary suitable for an embodiment of the present application. Figure 2 The example of the implementation illustrated in

[0060] Referring to Figure 2 , a plurality of facility areas (Z) can be provided in the working boundary 100, and the same or different types of facilities 120 can be provided in each of the facility areas (Z). The logistics robots 110 can pass through at least one or a portion of the plurality of facility areas (Z) while moving in the working boundary 100, and a plurality of logistics robots 110 can be provided in the working boundary 100.

[0061] During movement of the logistics robots 110 within the working boundary 100, collisions between the logistics robots 110, or between the logistics robots 110 and the facilities 120, can occur. Accordingly, when the logistics robots 110 collide, the logistics robots 110 can be damaged and cause a loss, and the entire process can be stopped, resulting in a decrease in process efficiency. Therefore, it is proposed to predict and prevent such collisions of the logistics robots 110 in advance by the collision prediction device 150 according to an embodiment of the present application.

[0062] Hereinafter, a configuration of a control device 140 suitable for an embodiment of the present application will be described with reference to Figure 3 An example of an implementation of the working boundary 100 suitable for an embodiment of the present application will be described.

[0063] Figure 3 is a block diagram illustrating an example of a configuration of a control device suitable for an embodiment of the present application. Figure 3 Each of the elements illustrated in

[0064] Referring to Figure 3The control device 140 can include a firmware management section 141, a traffic control section 142, a process management section 143, a production / logistics management section 144, an inventory management section 145, a communication section 146, a monitoring section 147, a map management section 148, and a task scheduling management section 149.

[0065] The firmware management section 141 can obtain the latest firmware of the logistics robot 110 through the communication section 146, and transmit the latest firmware to the logistics robot 110 to update the firmware, thereby keeping the firmware of the logistics robot 110 up to date.

[0066] The traffic control section 142 can control traffic lights and automatic doors based on the path of the logistics robot 110, and can recalculate the path of the logistics robot 110 according to the traffic situation.

[0067] The process management section 143 can define the process of each product, and can manage tasks such as process progress and progress location.

[0068] The production / logistics management section 144 can schedule the logistics robot 110 based on the task.

[0069] The inventory management section 145 can manage the location and quantity of each material, which is useful for more efficient process operation, for example, dispatching the logistics robot 110 to the destination for pallet pickup or collection in advance before detecting the point in time when the material is actually assembled / consumed.

[0070] The communication section 146 can communicate with internal elements of the work boundary 100, such as the logistics robot 110, the facility 120, and the detection device 130, and external entities such as a firmware update server.

[0071] The monitoring section 147 can monitor the location, path, battery status, communication status, and power system status of each individual logistics robot 110. Herein, the path is a concept including a global path based on landmarks and a local path in real time. In addition, the battery status can include voltage, current, temperature, peak of voltage and current, state of charge (SOC), and state of health (SOH). The communication status can include information on the currently activated communication protocol such as Wi-Fi, the connected AP, the distance to the AP, and the used channel. In addition, the power system status can include the load, temperature, and RPM of the drive system.

[0072] In addition, the monitoring section 147 can identify the currently assigned task, the working mode, and the firmware version of each individual logistics robot 110.

[0073] The map management section 148 can obtain map data in the form of a grid map obtained during travel of the AMR in the logistics robot 110 within the work boundary 100, and the map management section 148 can provide a tool for the factory manager to edit the obtained map data. By editing the map data, it is possible to set an area in which the logistics robot 110 enters to perform one or more preset works, a virtual lane, an intersection, and a restricted entry area, but these are merely examples and are not limited thereto. In addition, the map management section 148 can distribute the grid map initially obtained by the logistics robot 110 by actually traveling to other logistics robots 110 through the communication section 146.

[0074] The task scheduling management section 149 can manage and monitor the tasks of the logistics robots 110 based on the process information of the work boundary 100 received from the facility 120 and the detection device 130 through the communication section 146. In addition, the task scheduling management section 149 can select a specific logistics robot 110 and assign a task thereto, and set a global path of the logistics robot 110 according to the assigned task.

[0075] Next, the logistics robot will be described with reference to Figure 4

[0076] Figure 4 is a block diagram showing a configuration example of a logistics robot suitable for the embodiment of the present application.

[0077] Referring to Figure 4 , the logistics robot 110 can include a driving section 111, a sensing section 112, a loading section 113, a communication section 114, and a control section 115. Hereinafter, each element will be described.

[0078] The driving section 111 can include a driving source participating in movement, turning, and stopping of the logistics robot 110, a wheel, and a suspension. As the driving source, a motor receiving power from a built-in battery (not shown) can be used. The wheel can include one or more drive wheels receiving driving force from the driving source, and a non-drive wheel rotating by movement of the vehicle body without receiving driving force. According to the embodiment, when a plurality of drive wheels are provided, the driving source can be matched with each drive wheel, so that rotation of each drive wheel can be independently controlled. In this case, by making the rotation directions of different drive wheels different from each other, turning by rotating the vehicle body can be achieved without a separate turning device. At least a part of the non-drive wheel can be configured as a roller type wheel, but this is merely an example and is not limited thereto.

[0079] ​The sensing section 112 is used to detect the surrounding environment or working state of the logistics robot 110, and can include at least one selected from the following group: a two-dimensional laser scanner (e.g., LiDAR), a three-dimensional vision (stereo) camera, a multi-axis gyro sensor, an acceleration sensor, a wheel encoder, and a proximity sensor.

[0080] The encoder can output information for determining the amount of wheel rotation by utilizing light emitted from a light emitting element (e.g., a photodiode). For example, the encoder can count the number of slits provided in the circumferential direction on a disc that rotates along with the wheel or the wheel per unit time. The control section 115 can estimate displacement by analyzing the change in position over time based on data obtained through the encoder and the gyro sensor, thereby performing odometry. However, due to wheel slip or wear (the dynamic radius of the wheel can change), there can be an error between the actual displacement and the displacement estimated based on the encoder data. Accordingly, in performing odometry, the control section 115 can utilize a specific algorithm (e.g., an Extended Kalman Filter (EKF)) to perform correction of noise and error based on information collected from the wheel and the gyro sensor, thereby outputting a result having a tendency closer to the actual value. This odometry is particularly useful when current position determination (localization) cannot be achieved using the two-dimensional laser scanner described later.

[0081] The two-dimensional laser scanner can emit laser light around it by rotating a mirror and detect the reflected signal, thereby scanning the surrounding environment. Here, by analyzing the intensity of the reflected signal and the time difference between emission and reception, a detection result in the shape of a point cloud can be output.

[0082] The three-dimensional vision camera can calculate the distance to an object based on the parallax (i.e., the pixel distance between images obtained by each camera) between two cameras spaced apart from each other by a predetermined distance. Here, a texture projector for projecting infrared light having a specific pattern can be provided so that detection is possible even for planar objects having the same color (e.g., a white wall).

[0083] In general, the two-dimensional laser scanner can be used for mapping, navigation, and object recognition, and the three-dimensional camera can be used for obstacle avoidance during navigation. However, this is merely an example and is not necessarily limited thereto.

[0084] The loading section 113 is a device for loading objects to be transported, and can be implemented as a roof of a vehicle body, a table provided on the roof, an elevator, a turntable rotating along a vertical axis, a forklift, a conveyor, or a combination thereof. In the case of a forklift, a telescoping and tilting function similar to a forklift truck can be supported.

[0085] The communication section 114 can communicate with other elements such as the facilities 120, the control device 140 within the working boundary 100, can support communication between the logistics robots 110, and can communicate with the charging device when performing a charging task.

[0086] The control section 115 is a controller for performing overall control of the above-mentioned elements 111, 112, 113, and 114, and can perform determination of a current task, determination of a current position, determination of a destination, path planning, and control of a loading section based on information obtained from the control device 140 through the communication section 114.

[0087] Hereinafter, a process of predicting collision of a logistics robot according to an embodiment of the present application will be described with reference to Figure 5 A process of predicting collision of a logistics robot according to an embodiment of the present application will be described in detail.

[0088] Figure 5 is a sequence diagram illustrating a process of predicting collision of a logistics robot according to an embodiment of the present application.

[0089] Referring to Figure 5 The collision prediction device 150 according to an embodiment of the present application can include a collection section 151, a learning section 152, a prediction section 153, and an output section 154, and the collision prediction device 150 can be connected to the control device 140. However, the collision prediction device 150 according to an embodiment of the present application can include more or less elements than those illustrated in Figure 5 Main elements related to the description of the present application are illustrated, and more or less elements can be included in an actual implementation of the collision prediction device 150. In addition, the collision prediction device 150 can be implemented as a part of the control device 140, rather than being connected to the control device 140. Hereinafter, a process of collision prediction according to an embodiment of the present application will be described in detail.

[0090] First, the collection section 151 can collect process state information including facility information about a plurality of facilities 120 disposed in a preset working boundary 100 and logistics robot work information about work of a plurality of logistics robots 110 disposed in the working boundary 100 and moving through at least one or a part of a plurality of facility regions (Z) in which the plurality of facilities 120 are disposed, at step S510. In this case, the process state information can be obtained from the control device 140, and the control device 140 can obtain the logistics robot work information by communicating with the logistics robots 110 at step S501'. In addition, the control device 140 can obtain the facility information and the detection information by communicating with the facilities 120 and the detection device 130, and can transmit the facility information and the detection information to the collection section 151.

[0091] More specifically, the facility information can include at least one selected from a group of: coordinates of a facility area (Z) corresponding to the facility in the work boundary 100, and a task state of the facility. Specifically, the facility information can vary depending on the work characteristics of the facility 120. For example, when the facility 120 has the following work characteristics: the facility 120 is not fixed to a specific position in the facility area (Z) and the range occupied by the facility in the facility area (Z) varies depending on the task state, such a variation can be applied to the facility information, thereby improving the learning and judgment performance of the collision prediction.

[0092] Meanwhile, the logistics robot work information can include at least one selected from a group of: a speed, a moving direction, a moving path, a size, a normal work state, a sensor detection result, a braking specification, a real-time clock, a characteristic of an article loaded by each of the plurality of logistics robots 110, an inter-robot distance of some of the plurality of logistics robots, and a work priority order among the plurality of logistics robots.

[0093] In step S502, the collection part 151 can input the collected process state information to the learning part 512, and in step S503, the learning part 512 can learn a collision prediction criterion for each collision type based on the process state information, the collision type including a collision between any one of the plurality of facilities 120 and any one of the plurality of logistics robots 110, and a collision between some of the plurality of logistics robots. Subsequently, in step S504, the learning part 512 can input the learned collision prediction criterion to the prediction part 153.

[0094] Specifically, the learning part 152 can label the process state information according to a time period, and can learn a collision prediction criterion for each collision type based on the labeled process state information.

[0095] More specifically, for each collision type, the learning part 152 can classify the labeled process state information with respect to the plurality of logistics robots 110, according to whether a predetermined distance condition for a target object (including the facility 120 and another of the plurality of logistics robots 110) is satisfied. In this case, for example, when the logistics robot 110 approaches within a predetermined distance from the target object, the predetermined distance condition is satisfied.

[0096] In addition, the learning part 152 can further learn a collision prediction criterion based on process state information corresponding to a point in time before a predetermined time interval from a point in time at which the predetermined distance condition is satisfied in the labeled process state information. In this case, the learning of the collision prediction criterion can be performed when any one of the plurality of logistics robots 110 satisfies the predetermined distance condition.

[0097] Unlike this, the learning unit 152 can also learn the collision prediction criterion based on the process state information corresponding to the time point before the preset time interval from the time point at which the preset distance condition is not satisfied in the marked process state information.

[0098] In step S505, not only the learning information and the learning result but also the process state information as the determination information for predicting the collision can be continuously supplied to the prediction unit 153. Based on the learned collision prediction criterion and the process state information collected after the learning, the prediction unit 153 can predict the occurrence of the collision of at least one of the plurality of logistics robots 110.

[0099] Subsequently, in step S506, the prediction unit 153 can input the prediction result to the output unit 154, and in step S507, the output unit 154 can output the control signal corresponding to the input prediction result. In this case, the output control signal can be transmitted to the control device 140 and can be transmitted to the logistics robots 110 or the facility 120, or can be directly transmitted to the logistics robots 110 and the facility 120.

[0100] More specifically, the control signal can include a control signal for expressing a warning notification based on the prediction result, and the warning notification can be transmitted to the control device 140, the facility 120, and the logistics robots 110.

[0101] In addition, when the collision type corresponds to a collision between some of the plurality of logistics robots 110, the control signal can include a signal for controlling at least one of the speed or the moving direction or both of the speed and the moving direction of the logistics robots 110. Thereby, when the collision is predicted, the logistics robots 110 can be controlled to decelerate, stop, or perform an evasive travel. For example, the control signal can be a control signal for making the logistics robots stop in advance and wait, which can be transmitted to the logistics robots 110 via the control device 140 in step S508.

[0102] In addition, the control signal can control some of the logistics robots 110 so that the logistics robots work according to the order of work priorities between the logistics robots 110. For example, the logistics robots can be controlled so that the logistics robots 110 having a high priority maintain their working state, and the logistics robots 110 having a low priority can stop, decelerate, or perform an evasive travel.

[0103] Further, when the collision type corresponds to a collision between any one of the plurality of facilities 120 and any one of the plurality of logistics robots 110, a control signal can be transmitted to the facility 120, and the control signal can include a signal for controlling the task state of the facility 120. For example, by such a control signal, the facility 120 can be controlled to stop its task.

[0104] Further, when the collision type corresponds to a collision between any one of the plurality of facilities 120 and any one of the plurality of logistics robots 110, a control signal can be transmitted to the logistics robot 110, and the control signal can include a signal for controlling the speed or the moving direction or both of the logistics robot 110.

[0105] That is, the collision prediction device 150 can prevent the collision by controlling the operation of the logistics robot 110 or the facility 120 or both of the logistics robot 110 and the facility 120.

[0106] Further, after the collision problem related to the logistics robot 110 is resolved at step S509, the control device 140 can request the logistics robot 110 or the facility 120 to resume the task, and at step S510, the logistics robot 110 receiving the request can supplement its travel algorithm according to the result of the collision prediction. After the algorithm is supplemented, at step S511, the logistics robot 110 can request the control device 140 to reconfirm the task. After the task is confirmed, at step S512, the control device 140 can request the logistics robot 110 to continue the task operation, and finally, the logistics robot can resume the operation.

[0107] According to the above-described embodiments of the present application, the collision of the logistics robot can be predicted in advance. Specifically, by learning the collision prediction criteria with respect to the collision type that can occur in the operation boundary, the performance of the collision prediction can be improved.

[0108] Further, the logistics robot or the facility is controlled according to the result of the collision prediction, thereby preventing the collision from occurring.

[0109] Through the above-described effects, the logistics robot or the facility can be prevented from being damaged, and further, the process delay due to the occurrence of the collision event can be prevented, thereby enabling the process to be continued in a stable manner.

[0110] Although specific embodiments of the present application are described for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and alternatives are possible without departing from the technical idea of the present application as disclosed in the appended claims.

Claims

1. A collision prediction device for a logistics robot, the collision prediction device comprising: The collection unit is configured to collect process status information, which includes facility information about multiple facilities set in a preset work boundary and logistics robot work information about multiple logistics robots set in the work boundary and moving through at least one or a portion of multiple facility areas where multiple facilities are set. The learning department is configured to learn collision prediction standards based on process status information for various collision types, including collisions between any one of multiple facilities and any one of multiple logistics robots, as well as collisions between multiple logistics robots. The prediction unit is configured to: predict the occurrence of a collision between at least one of multiple logistics robots based on process state information collected after learning collision prediction criteria and the learned collision prediction criteria; and The output section is configured to output a control signal corresponding to the prediction result.

2. The collision prediction device for logistics robots according to claim 1, wherein, The learning unit is configured to: mark process status information according to time periods, and learn collision prediction standards for each collision type based on the marked process status information.

3. The collision prediction device for logistics robots according to claim 2, wherein, The learning unit is configured to classify the marked process status information for multiple logistics robots based on whether the preset distance conditions for the target object are met. The target object includes facilities and another of the multiple logistics robots.

4. The collision prediction device for a logistics robot according to claim 3, wherein, The learning unit is configured to: learn collision prediction criteria from the marked process status information based on the process status information corresponding to the time point before the preset time interval, starting from the time point when the preset distance condition is met.

5. The collision prediction device for a logistics robot according to claim 4, wherein, The learning unit is configured to perform learning based on any one of the multiple logistics robots meeting a preset distance condition.

6. The collision prediction device for a logistics robot according to claim 3, wherein, The learning unit is configured to: learn collision prediction criteria from the marked process status information based on the process status information corresponding to the time point before the preset time interval, starting from the time point when the preset distance condition is not met.

7. The collision prediction device for a logistics robot according to claim 1, wherein, The facility information includes at least one of the coordinates of the facility area corresponding to the facility in the work boundary or the task status of the facility.

8. The collision prediction device for a logistics robot according to claim 1, wherein, The facility information varies depending on the facility's operational characteristics.

9. The collision prediction device for a logistics robot according to claim 1, wherein, The operational information of the logistics robots includes at least one of the following: speed, direction of movement, path of movement, size, whether the operation is normal, sensor detection results, braking specifications, real-time clock, characteristics of each item loaded by multiple logistics robots, inter-robot distances between some of the multiple logistics robots, or the operational priority order among the multiple logistics robots.

10. The collision prediction device for a logistics robot according to claim 1, wherein, The process status information further includes sensor information measured by a detection device used to detect facilities and logistics robots within the work boundary.

11. The collision prediction device for a logistics robot according to claim 1, wherein, The control signals include control signals for enabling the expression of warning notifications based on prediction results.

12. The collision prediction device for a logistics robot according to claim 1, wherein, Based on the collision type corresponding to a collision between multiple logistics robots, the control signal includes a signal for controlling at least one of the speed or direction of movement of at least one of the multiple logistics robots.

13. The collision prediction device for a logistics robot according to claim 1, wherein, The control signals include signals for controlling multiple logistics robots so that the multiple logistics robots work according to the work priority order among the multiple logistics robots.

14. The collision prediction device for a logistics robot according to claim 1, wherein, Based on the collision type corresponding to a collision between any one of multiple facilities and any one of multiple logistics robots, the control signals include signals for transmitting to the facilities and controlling the task status of the facilities.

15. The collision prediction device for a logistics robot according to claim 1, wherein, Based on the collision type corresponding to a collision between any one of multiple facilities and any one of multiple logistics robots, the control signal includes a signal for transmitting to the logistics robot and controlling at least one of the logistics robot's speed or direction of movement.

16. A collision prediction method for a logistics robot, the collision prediction method comprising: Collect process status information, which includes facility information about multiple facilities set in a preset work boundary and logistics robot work information about multiple logistics robots set in the work boundary and moving through at least one or a part of multiple facility areas set with multiple facilities. Based on process status information, collision prediction standards are learned for each collision type. Collision types include collisions between any one of multiple facilities and any one of multiple logistics robots, as well as collisions between multiple logistics robots. Based on the process state information collected after learning the collision prediction criteria, and the learned collision prediction criteria, the occurrence of a collision of at least one of multiple logistics robots is predicted; and Output control signals corresponding to the prediction results.

Images