Efficient path planning method based on scene understanding and AMR robot

Abstract

The invention provides an efficient path planning method based on scene understanding and an AMR robot, and the method comprises the steps: obtaining a standard transportation path of a material transportation task, carrying out the material transportation task according to the standard transportation path, determining an occurrence place of a traffic retardation event when the traffic retardation event is detected in a material transportation process, and carrying out the execution of the material transportation task according to the determined occurrence place. Performing congestion scene analysis on the occurrence place of the traffic blocking event, and judging whether the traffic blocking event is a path congestion event according to a congestion scene analysis result, the path congestion event being that the current transportation path cannot pass, the method comprises the following steps of: when an event that materials cannot be delivered to a target location within a time limit required range is detected, or the passing time is too long, and when the passing blocking event is a path blocking event, re-executing transportation path planning on the material transportation task, so that the reliability and timeliness of material transportation in a hospital can be fully ensured, and the transportation efficiency of the hospital is improved. And efficient and reliable material transportation capability support is provided for each functional module of a hospital.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of autonomous mobile robots, in particular to a high-efficiency path planning method based on scene understanding and an AMR robot. BACKGROUND

[0002] The distribution and transportation of hospital materials including medicines, consumables, specimens and medical waste are one of the key links to maintain the normal transportation of various functional modules of the hospital. The introduction of intelligent logistics systems has greatly improved the efficiency of the distribution and transportation of hospital materials, and has reduced the problems of distribution errors and transportation delays caused by human factors. AMR (Autonomous Mobile Robot) is a robot with autonomous mobility and autonomous obstacle avoidance capability. Due to its strong environmental adaptability, easy deployment and strong flexibility, it is widely used in short-distance transportation scenarios in the logistics field, and intelligent logistics systems also use AMR robots for hospital material distribution and transportation.

[0003] Due to the large number of functional modules in the hospital, the various functional modules such as the pharmacy, the ward, the operating room, the medical center, the disinfection center, the emergency department and the outpatient department are scattered and have different distances, and at the same time, most of the functional modules in the hospital are located in open public activity areas. Therefore, considering the deployment space and deployment cost, it is difficult to arrange a dedicated "separation of people and vehicles" transportation channel for the intelligent logistics system of the hospital. Therefore, the various functional modules in the hospital need to plan the materials needed for the next day in advance, so as to concentrate on transportation during the off-peak hours at night when there are fewer people. For some material requirements that are limited by storage conditions and cannot be transported in advance, or unplanned material requirements due to unexpected situations, human distribution is often used for transportation, which often leads to difficulties in arranging material transportation personnel, difficulty in ensuring the timeliness of material transportation, and mismatch between material types / quantities and material requirements. SUMMARY

[0004] The present application is based on the above problems, and proposes a high-efficiency path planning method based on scene understanding and an AMR robot, which can fully guarantee the reliability and timeliness of hospital material transportation, and provide high-efficiency and reliable material transportation support for various functional modules of the hospital.

[0005] Therefore, the first aspect of the present application proposes a high-efficiency path planning method based on scene understanding, comprising:

[0006] receiving a material transportation task distributed by a task management module of an intelligent logistics system;

[0007] acquire a standard transportation path of the material transportation task, so as to execute the material transportation task according to the standard transportation path, the standard transportation path being a pre-defined path corresponding to start and end locations of the material transportation task;

[0008] detect a traffic blockage event in real time during the material transportation process;

[0009] when the traffic blockage event is detected, determine a location of occurrence of the traffic blockage event;

[0010] perform a congestion scene analysis on the location of occurrence of the traffic blockage event;

[0011] determine whether the traffic blockage event is a path blockage event according to the congestion scene analysis result, the path blockage event being an event that a current transportation path cannot be passed through or a time for passing through the current transportation path is too long to deliver the material to a target location within a time limit;

[0012] when the traffic blockage event is the path blockage event, re-perform transportation path planning for the material transportation task.

[0013] Further, the step of performing the congestion scene analysis on the location of occurrence of the traffic blockage event specifically comprises:

[0014] determine a congestion-related scene element of the location of occurrence of the traffic blockage event, the congestion-related scene element being a scene element located at the location of occurrence of the traffic blockage event and causing the traffic blockage event to occur;

[0015] acquire state data of the congestion-related scene element through a remote sensing module of the intelligent logistics system;

[0016] use the state data of the congestion-related scene element to evaluate a congestion situation of the location of occurrence of the traffic blockage event.

[0017] Further, the step of determining the congestion-related scene element of the location of occurrence of the traffic blockage event specifically comprises:

[0018] acquire an environmental image in front of a current position;

[0019] parse a scene element in front of the current position from the environmental image to generate a scene element list in front of the current position;

[0020] match each scene element in the scene element list with a congestion-related scene element in a congestion-related scene element database;

[0021] determine a congestion-related scene element in the scene element list according to a matching result.

[0022] Further, the step of evaluating the congestion condition of the occurrence location of the traffic jam event using the state data of the congestion-related scene elements specifically comprises:

[0023] identifying the scene type of the traffic jam event based on the environment image;

[0024] extracting, from the state data of the congestion-related scene elements, a congestion-related parameter corresponding to the scene type of each congestion-related scene element;

[0025] loading a congestion model corresponding to the traffic jam event to calculate the passability and / or waiting time at the occurrence location of the traffic jam event according to the congestion-related parameter.

[0026] Further, the step of determining whether the traffic jam event is a path congestion event according to the congestion scene analysis result specifically comprises:

[0027] when the value of the passability at the occurrence location of the traffic jam event calculated by the congestion model corresponding to the traffic jam event is less than a preset first threshold value, or the waiting time is greater than a preset second threshold value, determining that the traffic jam event is a path congestion event.

[0028] Further, the step of re-executing the transportation path planning of the material transportation task according to the congestion scene analysis result specifically comprises:

[0029] generating at least one alternative transportation path, the alternative transportation path being a transportation path with the current location as the starting point and the destination of the material transportation task as the ending point;

[0030] determining a jam-prone passing point on the alternative transportation path, the jam-prone passing point being a high-frequency occurrence location of a path congestion event;

[0031] performing real-time scene analysis on the jam-prone passing point;

[0032] determining one of the alternative transportation paths as a target transportation path according to the real-time scene analysis result;

[0033] configuring the target transportation path as a new transportation path to execute the material transportation task.

[0034] Further, after the step of determining whether the traffic jam event is a path congestion event according to the congestion scene analysis result, the method further comprises:

[0035] when the traffic jam event is a path congestion event, obtaining the coordinates of the occurrence location of the traffic jam event;

[0036] reporting the coordinates of the occurrence location of the traffic jam event to a server.

[0037] The step of determining the easily congested passing points on the alternative transportation path specifically comprises:

[0038] sending the alternative transportation path to a server;

[0039] receiving the easily congested passing points of the alternative transportation path returned by the server.

[0040] Further, the step of performing real-time scene analysis on the easily congested passing points specifically comprises:

[0041] determining congestion-related scene elements of the easily congested passing points;

[0042] obtaining state data of the congestion-related scene elements by a remote sensing module of the intelligent logistics system;

[0043] using the state data of the congestion-related scene elements to evaluate the congestion situation of the easily congested passing points.

[0044] Further, the step of determining one of the alternative transportation paths as the target transportation path according to the real-time scene analysis result specifically comprises:

[0045] traversing each of the alternative transportation paths to determine whether there is an alternative transportation path without an easily congested passing point that is currently experiencing a path congestion event;

[0046] when there are multiple alternative transportation paths without an easily congested passing point that is currently experiencing a path congestion event, calculating the travel time of the alternative transportation path, the travel time being the time required to travel from the current location to the destination of the material transportation task along the alternative transportation path;

[0047] selecting the alternative transportation path with the shortest travel time as the target transportation path.

[0048] The second aspect of the present application proposes an AMR robot comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the scene understanding-based efficient path planning method according to any one of the first aspect of the present application.

[0049] The application provides a high-efficiency path planning method based on scene understanding and an AMR robot, which acquires a standard transportation path of a material transportation task, executes the material transportation task according to the standard transportation path, detects a traffic blockage event in a material transportation process, determines an occurrence position of the traffic blockage event, executes congestion scene analysis on the occurrence position of the traffic blockage event, judges whether the traffic blockage event is a path congestion event according to a congestion scene analysis result, the path congestion event is an event that a current transportation path cannot be passed through or a transportation time is too long to deliver materials to a target position within a time limit, and when the traffic blockage event is the path congestion event, re-executes transportation path planning for the material transportation task, so that the reliability and timeliness of in-hospital material transportation of a hospital can be fully ensured, and high-efficiency and reliable material transportation support can be provided for each functional module of the hospital. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 FIG. 1 is a flowchart of a high-efficiency path planning method based on scene understanding provided by an embodiment of the application. DETAILED DESCRIPTION

[0051] In order to more clearly understand the above-mentioned purpose, features and advantages of the application, the application will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that the embodiments of the application and the features in the embodiments can be combined with each other without conflict.

[0052] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from those described herein, therefore, the protection scope of the application is not limited by the specific embodiments disclosed below.

[0053] In the description of the present application, the term "a plurality of" refers to two or more, unless otherwise explicitly specified. The terms "upper", "lower", and the like indicate the orientation or positional relationship shown in the drawings based on the orientation or positional relationship shown in the drawings, and are merely used for the purpose of facilitating the description of the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. The terms "connected", "mounted", "fixed", and the like should be interpreted broadly, for example, "connected" can be fixed connection, or detachable connection, or integral connection; can be directly connected, or indirectly connected through an intermediate medium. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. In addition, the terms "first", "second", and the like are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features limited by "first", "second", etc. can be explicitly or implicitly included one or more features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0054] In the description of the present application, the terms "one embodiment", "some embodiments", "a specific embodiment", and the like mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0055] A high-efficiency path planning method based on scene understanding and an AMR robot according to some embodiments of the present application are described below with reference to the accompanying drawings.

[0056] The high-efficiency path planning method based on scene understanding and the AMR robot provided by the present application are applied to a smart logistics system. The smart logistics system referred to by the present application refers to a software system for managing the transportation of materials in a hospital running on a server (which can be a cloud server or a local server deployed in the hospital). The software system manages logistics robots such as AMRs in the form of a network service through the Internet or a local network, including material transportation task allocation, logistics robot scheduling, and real-time monitoring of material transportation tasks.

[0057] The intelligent logistics system comprises a task management module and a remote sensing module. The task management module of the intelligent logistics system is a module for managing material transportation tasks in the intelligent logistics system. The material transportation task is a transportation task of transporting medical supplies from one location in a hospital to another location in the hospital according to real-time or pre-planned material requirements. The remote sensing module of the intelligent logistics system is a module for acquiring remote sensing data of various locations in the hospital.

[0058] More specifically, the server is communicatively connected with various sensors including visual sensors, infrared sensors, temperature sensors, etc. and electrical equipment such as elevators, access control, etc. set in the hospital through a local network or the Internet to acquire real-time remote sensing data of various locations in the hospital. These remote sensing data include visual data and state data corresponding to each sensor and electrical equipment.

[0059] As shown in Figure 1 The first aspect of the present application proposes an efficient path planning method based on scene understanding, comprising:

[0060] Receiving a material transportation task assigned by the task management module of the intelligent logistics system;

[0061] Acquiring a standard transportation path of the material transportation task to execute the material transportation task according to the standard transportation path. The standard transportation path is a pre-defined path corresponding to the start and end locations of the material transportation task;

[0062] Detecting a traffic blockage event in real time during the material transportation process;

[0063] When the traffic blockage event is detected, determining the occurrence location of the traffic blockage event;

[0064] Performing a congestion scene analysis on the occurrence location of the traffic blockage event;

[0065] According to the congestion scene analysis result, judging whether the traffic blockage event is a path congestion event. The path congestion event is an event that the current transportation path cannot be passed, or the passing time is too long to deliver the material to the target location within the time limit requirement;

[0066] When the traffic blockage event is a path congestion event, re-executing the transportation path planning for the material transportation task.

[0067] Specifically, the start and end points refer to the combination of the starting point and the target point of the material transportation task, each material transportation task has a starting point and a target point, which are the positions of a functional module in the hospital. In the technical solutions of some embodiments of the application, a standard transportation path is defined between the positions of each two functional modules in the hospital, such as a standard transportation path defined between the pharmacy and the ward, a standard transportation path defined between the consumable warehouse and the operating room, etc., which can be the shortest path between the two positions, or the shortest path in time between the two positions, etc. In the technical solutions of some other embodiments of the application, standard transportation paths can also be defined only between the positions of some key functional modules, and no standard transportation path is defined between two functional modules that do not generate material transportation tasks or generate material transportation tasks with very low probability, and the logistics robot or the server performs real-time path planning when an accidental material transportation demand occurs.

[0068] The passage blocking event is an event encountered by the AMR robot during passage, which requires the AMR robot to stop and wait, such as an event that the passage path is blocked by pedestrians or other obstacles, an event that there is no bypass path within the visual range of the AMR robot, or an event that the AMR robot travels to the elevator and needs to wait for passage, or an event that the AMR robot travels to the gate and needs to queue for passage, etc.

[0069] The path blocking event is one of the passage blocking events, which refers to the front of the planned current transportation path of the logistics robot being blocked, resulting in the current transportation path being unable to pass through or the transportation time being too long to deliver the material to the target point within the time limit. The reasons for path blocking are various, which can be that an obstacle completely blocks the passage of a passage point in the path, resulting in the passage point being unable to pass through and only being able to bypass or re-plan a new transportation path. It can also be that the current planned transportation path is crowded with people, resulting in a decrease in transportation efficiency and being unable to deliver the material to the target point within the time limit. This situation usually occurs at passage points such as elevators that need to be waited for.

[0070] Further, the step of performing a blocking scene analysis on the occurrence location of the passage blocking event specifically includes:

[0071] determining a blocking associated scene element of the occurrence location of the passage blocking event, the blocking associated scene element being a scene element located at the occurrence location of the passage blocking event that causes the occurrence of the passage blocking event;

[0072] obtaining state data of the blocking associated scene element through a remote sensing module of the intelligent logistics system;

[0073] The state data of the jam-related scene element is used to evaluate the jam situation at the occurrence location of the traffic blockage event.

[0074] Specifically, a complete scene is usually composed of multiple types of scene elements, including spatial elements such as architectural spatial forms, internal spatial layouts, spatial scales and proportions in a hospital, entity elements such as building components, furniture and decorations, and electrical equipment, and environmental elements such as sound elements, pedestrian elements, temperature elements, and odor elements. The instantaneous states of these elements together form an instantaneous scene, and the dynamic changes of a series of instantaneous scenes at a location over time constitute a dynamic scene of the location.

[0075] In the technical solution of the above embodiment, since the jam-related scene element is a scene element that causes the traffic blockage event to occur, the state data of the jam-related scene element refers to a set of values of state parameters of the jam-related scene element that is located at the occurrence location of the traffic blockage event in the traffic path of the AMR robot and affects the AMR robot traffic.

[0076] The state data of the jam-related scene element includes but is not limited to the personnel density at the occurrence location of the traffic blockage event, the waiting time of the elevator, the influence degree of the obstacle on the passability, etc., wherein the personnel density can be determined according to state parameters such as the size of the space at the occurrence location of the traffic blockage event and the number of personnel, the waiting time of the elevator can be determined according to state parameters such as the moving speed of the elevator, the moving direction of the elevator, the stay time of the elevator at each floor, the current floor of the elevator, and the floor at the occurrence location of the traffic blockage event, and the influence degree of the obstacle on the passability can be determined according to state parameters such as the position of the obstacle and the size of the obstacle.

[0077] Further, the step of determining the jam-related scene element at the occurrence location of the traffic blockage event specifically includes:

[0078] Obtaining an environmental image in front of the current position;

[0079] Parsing scene elements in front of the current position from the environmental image to generate a scene element list in front of the current position;

[0080] Matching each scene element in the scene element list with a jam-related scene element in a jam-related scene element database;

[0081] Determining the jam-related scene element in the scene element list according to the matching result.

[0082] In the technical solution of some embodiments of the present application, the image sensor such as the camera provided by the AMR robot can be used to obtain the environment image. The AMR robot can capture the environment image in front of its current position in real time during its travel, and the captured environment image during its travel can be directly used to identify the jam-related scene elements of the traffic blockage event. Of course, the remote sensing module of the intelligent logistics system can also be used to obtain the environment image in front of its current position. Since the cameras in the hospital are widely distributed and usually installed at a high place, their visual range is usually larger than that of the camera provided by the AMR robot, and the captured environment image can reflect more environment information. In the step of obtaining the environment image in front of the current position, the so-called current position refers to the occurrence location of the traffic blockage event.

[0083] In the technical solution of the above-mentioned embodiments, various scene elements that can cause the AMR robot to have a traffic blockage event in various scenes are configured in the jam-related scene element database, so that when the AMR robot has a traffic blockage event, it can match the scene elements in the environment image in front of the path of its current position to determine the potential scene elements causing the traffic blockage as the jam-related scene elements.

[0084] Further, the step of evaluating the jamming situation at the occurrence location of the traffic blockage event using the state data of the jam-related scene elements specifically includes:

[0085] identifying the scene type of the traffic blockage event based on the environment image;

[0086] extracting the jam-related parameters of each jam-related scene element corresponding to the scene type from the state data of the jam-related scene elements;

[0087] loading the jam model corresponding to the traffic blockage event to calculate the passability and / or waiting time at the occurrence location of the traffic blockage event according to the jam-related parameters.

[0088] The scene type of the traffic blockage event is the environment scene type at the occurrence location of the traffic blockage event, for example, it is an outdoor passageway, an indoor passageway, a waiting hall of an elevator, etc. According to actual implementation requirements, indoor and outdoor scenes can also be further subdivided, for example, a registration hall, a waiting hall, a building corridor, an outdoor parking lot, etc., so as to configure different jam-related parameters for different environment scene types.

[0089] The jam-related parameters are the parameters related to the passability of the AMR robot and the jam-related scene elements, such as the length and width of the passageway, the number of pedestrians, the current position and running direction of the elevator, the size of the obstacle, etc.

[0090] The jam model is a mathematical model configured for various traffic jam events to calculate the degree of jam, that is, a mathematical function of the passability and / or waiting time at the occurrence location of the traffic jam event, taking the jam-related parameters extracted from the state data of the jam-related scene elements at the occurrence location of the traffic jam event as calculation parameters. Different traffic jam events correspond to different jam models. For example, when the traffic jam event is a crowd-induced jam event, it corresponds to a crowd jam model that calculates the passability using the passage size and the crowd number sign; when the traffic jam event is an elevator waiting-induced jam event, it corresponds to an elevator waiting model that calculates the waiting time using the position, direction, and number of people waiting for an elevator on each floor of the elevator.

[0091] Further, the step of determining whether the traffic jam event is a path jam event according to the jam scene analysis result specifically includes:

[0092] When the numerical value of the passability at the occurrence location of the traffic jam event calculated by the jam model corresponding to the traffic jam event is less than a preset first threshold value, or the waiting time is greater than a preset second threshold value, it is determined that the traffic jam event is a path jam event.

[0093] As described above, the jam model is a mathematical function that quantitatively calculates the passability or traffic waiting time at the occurrence location of the traffic jam event according to the state data of the jam-related scene elements at the occurrence location of the traffic jam event. The passability at the occurrence location of the traffic jam event is reflected by a specific numerical value. The larger the numerical value, the better the passability. The smaller the numerical value, the worse the passability.

[0094] Further, the step of re-executing the transportation path planning of the material transportation task according to the jam scene analysis result specifically includes:

[0095] Generating at least one alternative transportation path, the alternative transportation path being a transportation path with the current location as the starting point and the destination of the material transportation task as the ending point;

[0096] Determining the jam-prone passing points on the alternative transportation path, the jam-prone passing points being high-frequency occurrence locations of path jam events;

[0097] Performing real-time scene analysis on the jam-prone passing points;

[0098] Determining one of the alternative transportation paths as the target transportation path according to the real-time scene analysis result;

[0099] Configuring the target transportation path as a new transportation path to execute the material transportation task.

[0100] The alternative transportation path is a transportation path starting from the current position of the AMR robot and ending at the destination of the material transportation task, and the alternative transportation path does not completely overlap with the remaining part of the current path. More specifically, the alternative transportation path is an alternative path that avoids the current path blockage event.

[0101] The easily blocked passing point is a high-frequency location of the path blockage event in the hospital, which is usually located in a densely populated area, a material storage area, or an elevator waiting hall, etc.

[0102] In the step of real-time scene analysis of the easily blocked passing point, the current environment scene of each easily blocked passing point on the alternative transportation path is analyzed to determine whether a path blockage event is occurring at each easily blocked passing point on the alternative transportation path.

[0103] In the step of determining one of the alternative transportation paths as the target transportation path according to the real-time scene analysis result, among the alternative transportation paths in which no easily blocked passing point is experiencing a path blockage event, the transportation path with the shortest travel time is selected as the target transportation path.

[0104] Further, after the step of determining whether the traffic jam event is a path blockage event according to the blockage scene analysis result, the method further comprises:

[0105] When the traffic jam event is a path blockage event, obtaining the coordinates of the location where the traffic jam event occurs;

[0106] Reporting the coordinates of the location where the traffic jam event occurs to the server;

[0107] The step of determining the easily blocked passing point on the alternative transportation path specifically comprises:

[0108] Sending the alternative transportation path to the server;

[0109] Receiving the easily blocked passing point of the alternative transportation path returned by the server.

[0110] In the technical solution of the above embodiment, the AMR robot generates the alternative transportation path by itself, that is, the AMR robot stores a global map of the hospital, which can be scanned by the AMR robot itself or issued to the AMR robot by the server, and generates at least one alternative transportation path based on the global map of the hospital.

[0111] In the technical solution of some other embodiments of the present application, the alternative transport path can also be generated by the server and then sent to the AMR robot, that is, the server selects the transport path with the shortest travel time and without the easily blocked passing point from the several alternative transport paths generated according to the current position of the AMR robot and the destination of the material transport task, and sends the selected transport path to the AMR robot as the target transport path.

[0112] Further, the step of performing real-time scene analysis on the easily blocked passing point specifically includes:

[0113] determining a jam-related scene element of the easily blocked passing point;

[0114] obtaining state data of the jam-related scene element by a remote sensing module of the intelligent logistics system;

[0115] evaluating the jamming situation of the easily blocked passing point using the state data of the jam-related scene element.

[0116] Similarly, in the technical solution of the above embodiment, after the step of performing real-time scene analysis on the easily blocked passing point, it is determined whether the easily blocked passing point is currently experiencing a path jamming event according to the result of evaluating the jamming situation of the easily blocked passing point.

[0117] Further, the step of determining one of the alternative transport paths as the target transport path according to the result of real-time scene analysis specifically includes:

[0118] traversing each of the alternative transport paths to determine whether there is an alternative transport path without the easily blocked passing point currently experiencing a path jamming event;

[0119] When there are multiple alternative transport paths without the easily blocked passing point currently experiencing a path jamming event, the travel time of the alternative transport path is calculated, which is the time required to travel from the current position to the destination of the material transport task along the alternative transport path;

[0120] selecting the alternative transport path with the shortest travel time as the target transport path.

[0121] Further, before the step of determining one of the alternative transport paths as the target transport path according to the result of real-time scene analysis, it further includes:

[0122] determining whether the passability of the current transport path is less than a preset third threshold value;

[0123] If the passability of the current transport path is less than the preset third threshold value, the step of determining one of the alternative transport paths as the target transport path according to the result of real-time scene analysis is performed.

[0124] If the passability of the current transportation path is greater than the preset third threshold value, the passing time of the current transportation path and the alternative transportation path is compared, and one of the transportation paths with the shortest passing time is determined as the target transportation path.

[0125] In the technical solution of the above embodiment, if the passability of the current transportation path is less than the preset first threshold value but greater than the preset third threshold value, the current transportation path is still used as the target transportation path when the passing time of the current transportation path is less than the passing time of the alternative transportation path.

[0126] The second aspect of the present application provides an AMR robot comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the scene understanding-based efficient path planning method according to any one of the first aspect of the present application.

[0127] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or device. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0128] In accordance with the embodiments of the present application as described above, these embodiments do not exhaustively describe all the details, nor limit the present application to only the specific embodiments described. Obviously, many modifications and variations can be made according to the above description. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can well utilize the present application and make modifications and uses based on the present application. The present application is limited only by the claims and their full scope and equivalents.

Claims

1. An efficient path planning method based on scene understanding, characterized in that: include: Receive material transportation tasks assigned by the task management module of the smart logistics system; Obtaining a standard transport path for the material transport task, so as to execute the material transport task according to the standard transport path, wherein the standard transport path is a predefined path corresponding to the starting and ending locations of the material transport task; Real-time detection of traffic obstruction events during material transportation; When a traffic obstruction event is detected, determining the location where the traffic obstruction event occurs; Performing a congestion scenario analysis on the location where the traffic blocking event occurs; Determine whether the traffic obstruction event is a path congestion event based on the congestion scenario analysis results. A path congestion event is an event in which the current transportation route is impassable or the travel time is too long to deliver the materials to the destination within the required time limit. When the traffic blocking event is a path blocking event, the transportation path planning for the material transportation task is re-executed.

2. The efficient path planning method based on scene understanding according to claim 1 is characterized in that: The step of performing a congestion scenario analysis on the location where the traffic blocking event occurs specifically includes: Determining a congestion-related scene element at the location where the traffic blocking event occurs, wherein the congestion-related scene element is a scene element located at the location where the traffic blocking event occurs and causing the traffic blocking event to occur; Acquiring the status data of the congestion-related scene elements through the remote sensing module of the smart logistics system; The status data of the congestion-related scene elements are used to evaluate the congestion situation at the location where the traffic blocking event occurs.

3. The efficient path planning method based on scene understanding according to claim 2 is characterized in that: The step of determining the congestion-related scene elements at the location where the traffic blocking event occurs specifically includes: Get the environment image in front of the current position; Analyzing the scene elements in front of the current position from the environmental image to generate a list of scene elements in front of the current position; matching each scene element in the scene element list with a congestion-related scene element in a congestion-related scene element database; Determine the blockage-related scene elements in the scene element list according to the matching result.

4. The efficient path planning method based on scene understanding according to claim 3 is characterized in that: The step of evaluating the congestion situation at the location where the traffic blocking event occurs using the status data of the congestion-related scene elements specifically includes: Identifying a scene type of the traffic blocking event based on the environmental image; extracting a congestion-related parameter corresponding to the scene type for each congestion-related scene element from the state data of the congestion-related scene element; A congestion model corresponding to the traffic blocking event is loaded to calculate the passability and / or waiting time at the location where the traffic blocking event occurs according to the congestion-related parameters.

5. The efficient path planning method based on scene understanding according to claim 4 is characterized in that: The step of determining whether the traffic blocking event is a path blocking event according to the congestion scene analysis result specifically includes: When the value of the accessibility at the occurrence location of the traffic blocking event calculated by the congestion model corresponding to the traffic blocking event is less than a preset first threshold, or the waiting time is greater than a preset second threshold, the traffic blocking event is determined to be a path congestion event.

6. The efficient path planning method based on scene understanding according to claim 2, characterized in that: The step of re-executing the transportation route planning for the material transportation task according to the congestion scenario analysis result specifically includes: Generate at least one alternative transport route, where the alternative transport route is a transport route starting from the current location and ending at the destination of the material transport task; Determine a waypoint on the alternative transport route that is prone to congestion, where the waypoint is a location where congestion events frequently occur; Performing real-time scene analysis on the congested waypoints; Determine one of the alternative transport routes as the target transport route based on the real-time scenario analysis results; The target transport path is configured as a new transport path to execute the material transport task.

7. The efficient path planning method based on scene understanding according to claim 6 is characterized in that: After determining whether the traffic blocking event is a path blocking event based on the congestion scenario analysis result, the method further includes: When the traffic blocking event is a path blocking event, obtaining the coordinates of the location where the traffic blocking event occurs; Reporting the coordinates of the location where the traffic blocking event occurs to the server; The step of determining the easily blocked waypoints on the alternative transport route specifically includes: Sending the alternative transport path to a server; The easily blocked waypoints of the alternative transport route returned by the receiving server.

8. The efficient path planning method based on scene understanding according to claim 6 is characterized in that: The step of performing real-time scene analysis on the congestion-prone waypoint specifically includes: Determining congestion-related scene elements of the congestion-prone waypoint; Acquiring the status data of the congestion-related scene elements through the remote sensing module of the smart logistics system; The congestion condition of the congestion-prone waypoint is evaluated using the status data of the congestion-related scene elements.

9. The efficient path planning method based on scene understanding according to claim 6, characterized in that: The steps of determining one of the alternative transport routes as the target transport route based on the real-time scenario analysis results specifically include: Traversing each alternative transport route to determine whether there is an alternative transport route without a path congestion event occurring at a path congestion-prone waypoint; When there are multiple alternative transport routes that do not have any congestion-prone waypoints experiencing congestion events, calculate the travel time of the alternative transport routes, where the travel time is the time required to travel from the current location to the destination of the material transport task along the alternative transport routes; The alternative transport route with the shortest travel time is selected as the target transport route.

10. An AMR robot, characterized in that: It includes a memory and a processor, and the processor executes a computer program stored in the memory to implement the efficient path planning method based on scene understanding as described in any one of claims 1 to 9.

Images