Mobile robot

Abstract

To improve the accuracy of autonomous route setting by a mobile robot.SOLUTION: Travel means 103 travels to a destination according to a set route. Determination means 106 determines whether there are obstacles on the route, the road surface condition of the route, and the presence or absence of preset no-go zones on the route on the basis of images acquired by imaging means during travelling along the route. Route control means 107 determines whether the route needs to be modified on the basis of the determination results from the determination means.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to control of a robot's travel path. [Background technology]

[0002] There are robots that can travel within farmland to assist agricultural work. These robots are required to avoid obstacles while traveling.

[0003] Therefore, Patent Document 1 describes a route search device that detects obstacles (rocks, trees, water, grass, buildings, etc.) using sensors such as cameras, predicts changes in the detected obstacles over time, estimates the state of the obstacles at the date and time of route travel, and performs route search taking into account whether the obstacles will affect travel.

[0004] Patent Document 2 describes a driving area discrimination device that extracts geometric features (such as unevenness, curvature, and unevenness variance of nearby areas) from shape data measured by a sensor such as a laser range finder, and applies machine learning to the extracted features to determine ruts as driveable areas and other areas as impassable areas due to obstacles. Patent Document 2 also describes that grass growing in the center of a driving path is determined to be an "impassable area."

[0005] Patent Document 3 describes a driving assistance device that identifies obstacles in blind spots, classifies the identified obstacles into static and dynamic obstacles based on whether or not the center position of the identified obstacles changes over time, and evaluates the risk of collision with the dynamic obstacles.

[0006] Patent Document 4 describes a method of setting a high encounter probability for routes for which dynamic obstacle information has existed in the past, and selecting the route with the lowest encounter probability from among all routes. Patent Document 4 also describes a method of preparing maps including a map of static obstacles and a map of a self-travelable area excluding static obstacles, acquiring information on dynamic obstacles while traveling, and avoiding routes with dynamic obstacles when a dynamic obstacle is detected. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Publication No. 2019-8431

[0008] [Patent Document 2] Japanese Patent Application Laid-Open No. 2009-175932

[0009] [Patent Document 3] International Publication No. 2016 / 117060

[0010] [Patent Document 4] Japanese Patent Application Laid-Open No. 2014-209293 Summary of the Invention [Problem to be solved by the invention]

[0011] In the prior art described in Patent Documents 1 to 4, the robot does not simultaneously assess obstacles and road surface conditions when traveling. Therefore, for example, the robot may travel along a route with no obstacles but poor road surface conditions. As a result, the robot may become unable to travel along the route or its traveling mechanism may be damaged. Furthermore, for example, when a robot travels through farmland, even if there are no obstacles and the road surface is in good condition, it may enter a no-travel zone outside the work area, such as a neighboring property, or may leave the farmland and collide with a car or person, resulting in an accident.

[0012] An object of the present invention is to improve the accuracy of autonomous route setting by a traveling robot. [Means for solving the problem]

[0013] The present invention provides a traveling robot having a traveling means for traveling to a destination according to a set route, a photographing means, a determining means for determining, based on images acquired by the photographing means while traveling along the route, whether or not there are any obstacles on the route, the road surface condition of the route, and whether or not there are any pre-set no-travel areas on the route, and a route control means for determining, based on the determination result by the determining means, whether or not the route needs to be modified. [Effects of the Invention]

[0014] According to this invention, the safety of traveling by a traveling robot is improved. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a block diagram showing the configuration of a running robot according to an embodiment of the present invention; [Figure 2] FIG. 2 is a diagram schematically illustrating an example of map information according to the embodiment. [Figure 3] 4 is a flowchart showing the operation of the embodiment. [Figure 4] FIG. 10 is a diagram illustrating an example of operation of the embodiment. [Figure 5] FIG. 10 is a diagram illustrating an example of updated map information according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0017] Figure 1 is a block diagram showing the configuration of a mobile robot 1 according to one embodiment of the present invention. The mobile robot 1 is a robot that travels over farmland and performs agricultural tasks such as weeding and transporting harvested crops. Note that farmland can be any outdoor land used for growing crops, grazing livestock, etc., regardless of the type of crop grown, the level or method of management, etc.

[0018] As shown in FIG. 1, the mobile robot 1 has a processor 10 that controls the entire system, an input unit 21 for inputting various information, a display unit 22 such as an LCD panel, a drive unit 23 that drives various parts of the mobile robot 1, such as the wheels, a camera 30 that captures images of the area in front of the mobile robot 1, and a memory unit 40. The memory unit 40 consists of a volatile memory unit such as RAM and a non-volatile memory unit such as a hard disk. The volatile memory unit is used as a work area by the processor 10. The non-volatile memory unit stores map information 41, route information 42, and a program 48. The map information 41 is information that shows a map of the farmland where the mobile robot 1 will work.

[0019] This map information 41 divides the farmland to be managed into multiple areas, and for each area, defines the objects or conditions located in that area, such as obstacles, drivable roads, and poor road surface conditions (e.g., puddles). Route information 42 is information that indicates the travel route of the mobile robot 1 on the map shown by the map information 41. Program 48 is a program for controlling the mobile robot 1 to perform tasks related to farm work (such as transporting supplies and harvested products, inspecting crops and livestock, and spraying pesticides).

[0020] Figure 2 shows a schematic diagram of an example of map information 41. The land under management is divided into two-dimensional areas, and objects indicating the status of each area are superimposed. Object OB1 is an area where crops are planted and where mobile robots should not enter, object OB2 is an area where obstacles such as agricultural machinery exist, object AR1 is higher in elevation than the surrounding area (it is a hill), and therefore poses a certain degree of risk to travel, while the rest of the area is one where there is no particular obstacle to travel. Object PR1 indicates land outside of management.

[0021] In this embodiment, the processor 10 executes the program 48 to function as an acquisition means 101, a reception means 102, a driving means 103, a photographing means 104, an update means 105, a determination means 106, a route control means 107, and a setting means 108.

[0022] Here, the acquisition means 101 is a means for importing map information 41 from the outside via the input unit 21 and storing it in the storage unit 40. This acquisition means 101 may receive image data obtained by aerial photography of farmland via the input unit 21 and generate the map information 41 based on the received image data. The reception means 102 is a means for receiving settings of the starting point and destination of the traveling robot 1 on the map indicated by the map information 41.

[0023] The traveling means 103 is a means for controlling the traveling robot 1 to travel along a route indicated by the route information 42. Here, the route information 42 may be set by a human or may be searched for by the processor 10 of the traveling robot 1. That is, the traveling robot 1 may be a semi-autonomous type that is given information on part or all of the route to be traveled (destination, waypoints, etc.), or a fully autonomous type that determines by itself the necessary traveling plan for accomplishing a given task, including the destination. Note that even if the route information is specified in advance (i.e., before the start of traveling), the traveling robot 1 has a function of appropriately correcting the route based on images captured during traveling, as will be described later.

[0024] The imaging means 104 is a means for capturing images of the area in front of the traveling robot 1 using the camera 30. In this embodiment, the camera 30 is configured as a stereo camera (two cameras) capable of measuring depth. The imaging means 104 converts the image into a point cloud (pixel group). Each pixel has an object class and distance. This pixel group is generated using AI semantic segmentation and object localization techniques.

[0025] The update means 105 is a means for updating the map information 41 based on the image acquired by the photographing means 104. More specifically, if the image acquired by the photographing means 104 indicates that an obstacle has appeared in a certain area, information indicating the presence of the obstacle is mapped to that area in the map information 41. Furthermore, if the image acquired by the photographing means 104 indicates that a puddle has appeared in a certain area, information indicating the road surface condition of a puddle is mapped to that area in the map information 41.

[0026] The determination means 106 is a means for determining whether or not there are obstacles on the route, the road surface condition of the route, and whether or not there are any preset no-travel areas on the route, based on images acquired by the imaging means 104 while the traveling robot 1 is traveling along the route. This determination is made based on the map information 41 and the route information 42. That is, the determination means 106 determines, based on the map information 41, whether or not there are any obstacles, areas with poor road surface conditions, or preset no-travel areas along the route indicated by the route information 42.

[0027] Here, obstacles include people, ladders, agricultural equipment, baskets, fences, posts, rocks, fallen branches, etc. Road surface conditions include unevenness, slope, degree of mud, and other factors that affect the risk of driving. No-driving areas are areas where the robot is prohibited from driving even if there are no obstacles on the route and the road surface is in a condition that allows the robot to drive, such as farmland managed by others or public roads.

[0028] The route control means 107 is a means for determining whether or not the route needs to be corrected based on the determination result by the determination means 106 .

[0029] A specific example of determining whether a route change is necessary is as follows: If the set route (the route indicated by the route information 42) avoids static obstacles, does not enter prohibited areas, and travels on the best road surface conditions, it is determined that no route change is necessary; otherwise, it is determined that a route change is necessary. If an obstacle on the set route is a moving obstacle (such as a farmer or another robot), it is determined that no route change is necessary and the robot should wait. Whether a route change is necessary is also determined based on the nature of the moving obstacle and / or the existence or non-existence of an alternative route and the risk (cost) of the alternative route. If people or other robots are moving busily along the route randomly (i.e., unpredictably) and if a reasonable alternative route is available, it is determined that the route should be changed. This is because in such cases, it is estimated that the traveling robot 1 is likely to get in the way of people or other robots.

[0030] In addition, machine learning techniques may be used to determine at least one of the road surface conditions (whether it is a paved road or other area that is somewhat prepared for travel, an unprepared area such as grass, an area where crops are planted, or muddy, etc.) based on images taken by camera 30, determine the presence or absence and nature of obstacles, and determine no-drive zones.

[0031] The route control means 107 has a function of searching for an alternative route when it is determined that a route correction is necessary.

[0032] The setting means 108 is a means for searching for an optimal route from the departure point accepted by the accepting means 102 to the destination based on the current map information 41, and setting it as route information 42. The route information 42 obtained by the route control means 107 may include unnecessary routes, but the setting means 108 can obtain route information 42 that does not involve unnecessary routes from the departure point to the destination (in other words, is efficient).

[0033] Next, the operation of this embodiment will be described. Fig. 3 is a flowchart showing the operation of the processor 10. In this embodiment, the processor 10 repeatedly executes driving control (step S1), map information update (step S2), and route control (step S3) in accordance with the program 48.

[0034] In the travel control (step S1), the processor 10 controls the traveling robot 1 to travel along the route indicated by the route information 42. More specifically, the processor 10 estimates the current position of the traveling robot 1 by comparing the image acquired by the photographing means 104 with the map indicated by the map information 41. If the current position deviates from the route indicated by the route information 42, the processor 10 controls the current position to return to the route.

[0035] Next, in updating the map information (step S2), the processor 10 detects changes in objects on the map (e.g., the occurrence of a moving obstacle, a change in road surface conditions, etc.) based on images acquired by the photographing means 104, and reflects these changes in the map information 41.

[0036] Next, in path control (step S3), the processor 10 determines whether or not the path needs to be corrected based on the map information 41 and the path information 42, and if correction is necessary, corrects the path. This path correction is performed with the highest priority on the path closest to the current position of the traveling robot 1. Then, when step S3 is completed, the process returns to step S1.

[0037] FIG. 4 is a diagram showing an example of operation of this embodiment. FIG. 4 shows a portion of the map information 41 and route information 42 stored in the storage unit 40. In this operation example, the running robot 1 initially travels through an area sandwiched between prohibited areas 411 and 412. However, this road branches into two at prohibited area 413. Assume that the route of the running robot 1 was originally intended to pass through the area between prohibited areas 411 and 413. However, an image acquired by the imaging means 104 shows that a muddy area 421 caused by rain exists between prohibited areas 411 and 413. Therefore, the route control means 107 modifies the route of the running robot 1 to a route that passes through the area sandwiched between prohibited areas 413 and 412.

[0038] This will be explained in more detail using the example of FIG. FIG. 5 shows the map information of FIG. 3 that has been updated after the traveling robot 1 has traveled. In this example, it is specified to travel from the starting point PS to the destination point PE. Based on the latest map information before the update, and taking into account the terrain, etc., the walking robot 1 determines that the route TR1, which goes in a straight line from the starting point PS to the destination point PE, is the most risk-free and efficient route.

[0039] However, after setting off, as a result of taking photographs while driving, the robot recognizes the presence of a muddy area AR2 (an area with poor road conditions) at point P1, and determines that the risk of driving if it continues along the original route exceeds a predetermined level, so that the route setting needs to be revised (i.e., a new route needs to be searched for).As the robot continues to take photographs while driving, and searches for new routes as needed, such as detecting a new worker (object OB3) along the way, it ultimately searches for route TR2, which it determines to have the least risk, and reaches the original target value PE.

[0040] According to this embodiment, the mobile robot 1 can avoid obstacles, avoid no-travel areas, and travel on roads with better surface conditions. This improves the safety of travel. As a result, it is expected that the introduction of ground mobile robots in the agricultural field will be promoted.

[0041] Generally, even if a robot has the function of transmitting a GPS signal to determine its current location, there are often cases where the GPS does not function adequately in agricultural land, etc. In such cases, if the current location of the mobile robot cannot be determined (i.e., the mobile robot loses track of its own location), even if the latest and accurate map information is acquired, the map information cannot be effectively utilized. According to this embodiment, the current location of the mobile robot 1 is estimated by comparing the image acquired by the image capturing means 104 with the map information 41, so that the optimal route for the mobile robot 1 to travel can be determined even in an environment where GPS cannot be used. In addition, since farmland is generally vast, it may not be practical to send map data and control instructions to a mobile robot using a wireless communication device such as a Wi-Fi router. However, according to this embodiment, there is no need to send instructions or various information to the mobile robot 1 while it is moving, so the mobile robot 1 can move safely and reliably regardless of the size of the farmland or the communication environment.

[0042] Furthermore, in outdoor areas such as farmland that are likely to be exposed to natural influences, changes in the terrain occur frequently due to weather such as heavy rain (ground erosion due to rain, fallen trees due to typhoons, branches and greenhouses blown by wind, etc.), and therefore the map information provided to the mobile robot 1 may not reflect the current terrain or the locations of obstacles. There may also be cases where the status of obstacles changes drastically (such as when people or work machines move around within the farmland). In this embodiment, even in such cases, the mobile robot 1 can select and travel a reasonable route (a route that ensures a certain level of safety and does not take unnecessary detours).

[0043] <Other embodiments> Although one embodiment of the present invention has been described above, other embodiments of the present invention are also possible, for example as follows.

[0044] In the above embodiment, the path control means 107 of the mobile robot 1 determined whether a path correction was necessary and searched for an alternative path if a path correction was necessary. However, the path control means 107 may only determine whether a path correction is necessary, and an external server may search for an alternative path if a path correction is necessary. Specifically, the mobile robot 1 is equipped with a wireless communication module for wirelessly communicating with an external server that stores map information. In response to a request from the mobile robot 1, the external server determines an alternative path using a predetermined algorithm based on the location information received from the mobile robot 1 and the map information held by the server, and transmits the determined alternative path to the mobile robot 1.

[0045] The map information 41 may include only information about objects whose positions and shapes do not change over time, such as stationary obstacles, while information about objects that change over time, such as moving obstacles and road surface conditions, may be included in information separate from the map information 41. In this case, information about objects that change over time may be stored in a RAM or the like that can be accessed at high speed.

[0046] The setting means 108 may set a route based on information indicating the traveling performance of the traveling robot 1 (ground clearance, approach angle, departure angle, traction, suspension performance, drive system, tire performance, etc.). For example, suppose there is a limit value for the gradient of the road surface on which the traveling robot 1 can travel. In this case, the setting means 108 determines, as the traveling route of the traveling robot 1, a route that passes through only areas where the gradient of the road surface is smaller than the limit value.

[0047] In short, the traveling robot of the present invention only needs to have a traveling means for traveling to a destination according to a set route, a photographing means, a determination means for determining whether or not there are any obstacles on the route, the road surface condition of the route, and whether or not there are any pre-set no-travel areas on the route based on images acquired by the photographing means while traveling along the route, and a route control means for determining whether or not the route needs to be modified based on the determination result by the determination means. [Explanation of symbols]

[0048] 1...traveling robot, 10...processor, 21...input unit, 22...display unit, 23...driving unit, 30...camera, 40...memory unit, 41...map information, 42...route information, 48...program, 101...acquisition means, 102...reception means, 103...traveling means, 104...photographing means, 105...updating means, 106...determination means, 107...route control means, 108...setting means.

Claims

1. a travel means for traveling to a destination according to a set route; Photography means, a determining means for determining whether or not there is an obstacle on the route, the road surface condition of the route, and whether or not there is a predetermined no-travel area on the route, based on an image acquired by the photographing means while the vehicle is traveling along the route; a route control means for determining whether or not the route needs to be corrected based on the determination result by the determination means; A running robot having the above structure.

2. The vehicle further includes an update unit that updates map information used when determining the route based on the image captured by the image capturing unit. The running robot according to claim 1.

3. The map information includes at least information indicating the presence or absence of obstacles and road surface conditions for each divided area, When the determining means determines that the route needs to be corrected, the route control means determines an alternative route based on the map information. The traveling robot according to claim 2.

4. A reception means for receiving input of a departure point and a destination point; means for acquiring the map information; a setting means for setting a route from the departure point to the destination point based on the road surface condition; Further having The traveling robot according to claim 3 .

5. The setting means sets the route further based on information indicating the running performance of the running robot. The traveling robot according to claim 4.

6. A computer for controlling a robot equipped with a photographing means, a step of determining whether or not there is an obstacle on the route, the road surface condition of the route, and whether or not there is a predetermined no-travel area on the route, based on an image acquired by the photographing means while the vehicle is traveling along the set route; determining whether or not the route needs to be corrected based on the result of the determination; A program to execute.

Images