Autonomous mobile robot, program

Abstract

To provide an autonomous traveling robot that can appropriately avoid collision with a movable body approaching from the front.SOLUTION: An autonomous traveling robot 100 allows itself to autonomously travel along a travel route. When the autonomous traveling robot detects a specific movable body moving in front of the robot based on imaging data during the autonomous travel, it follows and detects the relative distance from the robot to the specific movable body, and performs a passing operation for the detected specific movable body in an operation mode according to the type of the specific movable body.SELECTED DRAWING: Figure 1

Description

[[Technical Field]]

[0001] The present invention relates to the technology of autonomously movable devices. [[Background Art]]

[0002] Patent Document 1 discloses that an obstacle around the own device is detected by a sensor, and the detection result is used for autonomous An autonomously traveling robot that travels is described. Further, in an autonomously traveling robot, the own device A travel route to a destination is calculated using detection results for obstacles around the device, and the calculated An autonomously traveling robot that travels according to a travel route is also known. [[Prior Art Documents]] [[Patent Documents]]

[0003] [[Patent Document 1]] Japanese Unexamined Patent Application Publication No. 2019-70952 [[Summary of the Invention]] [[Problem to be Solved by the Invention]]

[0004] Assume a situation where a moving object approaches an autonomously traveling robot while the autonomously traveling robot is traveling along an indoor passage or the like. In a configuration where the autonomously traveling robot travels along a passage or the like according to a calculated travel route, the travel route is calculated so as to avoid collision with the approaching moving object, and the vehicle travels according to the calculated travel route. However, while the autonomously traveling robot is traveling, the moving object may change its course to avoid collision with the autonomously traveling robot, and when both move in the same direction, the two may come close to each other without passing each other. the two may come close to each other without passing each other.

[0005] The present invention has been made in view of the above problems, and provides an appropriate collision avoidance with a moving object approaching from the front The objective is to provide an autonomous driving robot and program that can avoid obstacles. . [Means for solving the problem]

[0006] To solve the above problems, the autonomous mobile robot disclosed in this embodiment includes a sensor for detecting objects located around itself, an imaging unit for acquiring imaging data of the area in front of itself, a driving unit, an autonomous driving control unit that calculates a driving path in a path calculation area around itself based on the position of objects detected by the sensor, and drives the driving unit according to the calculated driving path to make the robot autonomously move, and an object detection unit that, during autonomous driving, detects a specific moving object located in front of the path calculation area relative to itself based on the imaging data, and determines the type of the detected specific moving object. The object detection unit is capable of determining the type of the specific moving object, The specific moving objects included in the imaging data are, in order, identified as: a transport device with a carrier, a person holding a walking assistance device, or a person in a wheelchair as a person requiring walking assistance; and a person not holding a walking assistance device and not in a wheelchair as a healthy person. The autonomous driving control unit controls the driving drive unit to perform a passing maneuver in which the autonomous driving unit and the detected specific mobile object pass each other, in an operating mode corresponding to the type of specific mobile object identified.

[0007] In the autonomous mobile robot with the above configuration, the robot calculates a travel path in the path calculation area ahead of itself based on the positions of surrounding objects detected by sensors, and drives the driving unit according to the calculated travel path, thereby enabling the robot to travel autonomously. During autonomous travel, if a specific moving object located in front of the robot is detected based on imaging data acquired by the imaging unit, The type of specific mobile entity is determined in the following order: transport equipment accompanied by a carrier, persons requiring walking assistance (people holding walking assistance devices or people in wheelchairs), and able-bodied individuals.The autonomous robot controls the driving unit to perform a passing maneuver, causing the autonomous robot to pass the specific mobile object in an operating mode appropriate to the type of specific mobile object identified. As a result, when a specific mobile object is approaching the autonomous robot from the front, the autonomous robot performs a passing maneuver in an operating mode appropriate to the type of specific mobile object, before the specific mobile object enters the path calculation area set around the autonomous robot. Therefore, after the specific mobile object enters the path calculation area, the autonomous robot can prevent the two from approaching each other without passing each other by avoiding each other in the same direction. Furthermore, by determining the type of specific mobile object in the order of transport equipment with a carrier, the person receiving walking assistance, and then a healthy person, it is possible to prevent the carrier included in "transport equipment with a carrier" from being mistakenly identified as a "healthy person."

[0008] Object detection unit but , as a specific moving object included in the imaging data, When the autonomous driving control unit detects a transport device with a person inside, it controls the driving unit to perform a passing maneuver, which involves moving the autonomous driving unit out of the way so as not to obstruct the path of the transport device with the person inside. In situations where a transport device with a person inside is approaching the autonomous driving robot, since the transport device with a person inside has less freedom to change its path than the autonomous driving robot, it is better for the autonomous driving robot to stop to the side of the passage so as not to obstruct the path of the transport device with the person inside, thereby allowing the autonomous driving robot to pass the transport device with the person inside. This allows the autonomous driving robot to perform an appropriate passing maneuver when a transport device with a person inside is approaching from the front.

[0009] Object detection unit but Furthermore, as specific moving objects included in the imaging data, a person holding a walking assistance device or a person in a wheelchair is identified as a person requiring walking assistance. If The autonomous driving control unit controls the driving unit to perform an avoidance maneuver, such as a passing maneuver, which causes the robot to travel along a path that avoids the path of the person receiving walking assistance. When the person receiving walking assistance is approaching the autonomous driving robot, the person is moving slower than the autonomous driving robot and has difficulty changing direction, so it is better for the autonomous driving robot to change direction and pass by the side of the person receiving walking assistance. This allows for the execution of an appropriate passing maneuver when the person receiving walking assistance is approaching from the front.

[0010] Object detection unit but , as a specific moving object included in the imaging data, When the autonomous driving control unit identifies a person who is not using a walking aid and is not in a wheelchair as a healthy person, it controls the driving unit to perform a passing maneuver, reducing the vehicle's speed to a slow speed. In situations where a healthy person is approaching the autonomous driving robot, it is better to allow the healthy person to pass the robot, as they can change their course more freely than the autonomous driving robot. This allows for the execution of an appropriate passing maneuver when a healthy person is approaching from the front. Walking aids include canes, walkers, rolling carts, and mobility scooters. Canes include white canes, T-shaped canes, multi-legged canes, crutches, and Lofstrand crutches.

[0011] The autonomous driving control unit, when a specific mobile object is moving in a direction that approaches its own device, The drive unit is controlled to perform a passing maneuver. In the above configuration, the vehicle approaches its own device. Only for a specific moving object moving in the opposite direction, the passing operation is started before this specific moving object enters the route calculation area . This makes it possible to suppress unnecessary execution of control for passing the specific moving object. [Effects of the Invention]

[0012] According to the present invention, it is possible to suppress a situation where an autonomous traveling robot and a specific moving object come close to each other without passing each other. [Brief Description of the Drawings]

[0013] [Figure 1] It is a perspective view of an autonomous traveling robot. [Figure 2] It is a configuration diagram of an autonomous traveling robot. [Figure 3] It is a flowchart explaining the procedure of autonomous traveling. [Figure 4] It is a diagram explaining a cost map. [Figure 5] It is a flowchart explaining the procedure of processing related to detection of a moving object. [Figure 6] It is a diagram explaining a specific moving object detected from imaged data. [Figure 7] It is a diagram explaining a range for tracking a specific moving object. [Figure 8] It is a flowchart explaining the process of S15 in FIG. 3. [Figure 9] It is a diagram explaining a slow traveling operation. [Figure 10] It is a diagram explaining a comparative example. [Figure 11] It is a diagram explaining an avoidance operation. [Figure 12] It is a diagram explaining a retraction operation. [Mode for Carrying Out the Invention]

[0014] (First Embodiment) The autonomous mobile robot according to this embodiment will be described with reference to the drawings. The device is intended for use indoors, such as in hospitals. (As shown in Figures 1 and 2) To that end, the autonomous mobile robot 100 consists of a housing 90, a controller 10, and a distance measuring sensor 1 1, imaging unit 12, memory 13, driving unit 14, display 15, and internal sensor It mainly features Sa16.

[0015] The housing 90 is a part that constitutes the external appearance of the autonomous mobile robot 100, and the upper housing 9 It consists of 1 and the lower housing 92. The upper front side of the upper housing 91 has a display I-15 is installed. Display 15 is a well-known touch panel, and control In response to instructions from roller 10, text and icons are displayed. In this embodiment, The display 15 is mounted on the upper front of the upper housing 91 so that the display surface faces forward. It is attached.

[0016] A stop button (not shown) is mounted on the top of the upper housing 91. The stop button is It is connected to controller 10, and when someone nearby operates the stop button, the controller It becomes possible to input a command to stop the autonomous mobile robot 100 to the roller 10. A push handle 94 is provided on the upper rear side of the side housing 91. Autonomous mobile robot 1 If 00 stops, a person nearby can operate the push handle 94 to launch the autonomous mobile robot 1 It is possible to move 00 manually.

[0017] Inside the housing 90 are the controller 10, the distance measuring sensor 11, the imaging unit 12, and memory. It houses 13 and the driving unit 14. The distance measuring sensor 11 is, for example, LiDAR ( Light Detection and Ranging (Laser Imaging Detection and Ranging) is used in the surrounding area. By emitting near-infrared and visible light and detecting the reflected light with an optical sensor, the device can respond to objects around it. The controller 10 acquires scan data that includes the measurement points. Each measurement point determines the relative distance from the device to surrounding objects, as well as the positional relationship between multiple measurement points. This makes it possible to determine the shape of the object. Below, the distance measuring sensor 11 determines the relative distance The information output by measurement is also referred to as scan data. Scan data includes the surrounding area. The two-dimensional measurement points on the surrounding object are associated with information indicating the emission direction of the distance measuring sensor 11. It is included in this configuration. In addition, the distance measuring sensor 11 is configured to emit light in a three-dimensional direction. If available, the scan data may include 3D measurement points.

[0018] The information indicating relative distance based on scan data is based on the position of the autonomous mobile robot 100. These are values ​​in the robot coordinate system, and therefore differ from the absolute coordinate system described later. The T coordinate system is defined as the direction of travel of the autonomous mobile robot 100 as "x", and the horizontal direction perpendicular to the direction of travel. This is a coordinate system where the direction is "y". Furthermore, the robot coordinate system includes the autonomous mobile robot 100. This may include information indicating posture (orientation).

[0019] The imaging unit 12 acquires imaging data obtained by imaging the area in front of the device. Specifically The imaging unit 12 is housed inside the upper housing 91 with its imaging direction facing forward. In this embodiment, the imaging unit 12 uses a monocular camera, but a compound camera can be used. It's okay to have it.

[0020] The lower housing 92 houses the drive unit 14. The drive unit 14 is autonomous This is the part that enables the mobile robot 100 to move, and consists of a pair of wheels 141A, 141B, It includes motors 142A and 142B and a motor driver 143. In this embodiment, The drive system used by the drive unit 14 is a differential two-wheel drive. In the differential two-wheel drive, as described later... The motor driver 143 is driven according to the command speed input from the controller. By distributing the driving force to motors 142A and 142B, the wheels 141A and 141B It is possible to perform straight-line and turning maneuvers. The driving unit 14 has two wheels It is not limited to having one wheel, and may have two or more wheels. Furthermore, the drive unit 14 It may be equipped with tracks or crawlers instead of wheels.

[0021] The internal sensor 16 detects the posture of the autonomous mobile robot 100 as it moves using the driving unit 14. (Direction) and odometry indicating the amount of movement of the wheels 141A and 141B provided by the drive unit 14. It consists of sensors that detect [something].

[0022] Memory 13 contains an environmental map 30 corresponding to the indoor environment in which the autonomous robot 100 is operating, and The program 20 that the controller 10 will run is stored. The environmental map 30 shows indoors The positions of passages, walls, stairs, and other obstacles are stored in an absolute coordinate system W. Furthermore, the environmental map 30 identifies the type of each obstacle in relation to indoor obstacles. The tag information and the driving speed that the autonomous robot 100 must maintain may also be stored.

[0023] Next, the configuration of controller 10 will be described. Controller 10 includes a CPU (not shown), volatile It includes static memory and volatile memory. In controller 10, the CPU uses memory 13. By executing the stored program 20, the measurement data processing unit 21 and the path calculation unit 2 2. Coordinate transformation unit 23, self-position estimation unit 24, state management unit 25, drive control unit 26, object detection It functions as part 27, passing motion determination unit 28, and display control unit 29. In addition to the above, Some of the components may be implemented by the hardware provided by the controller 10. In this embodiment, the route calculation unit 22, the state management unit 25, and the drive control unit 26 are used for autonomous driving. This is an example of a control unit.

[0024] Next, using Figure 3, we will explain the control for autonomous driving performed by the controller 10. Each process shown in Figure 3 is executed by each function of the controller 10 shown in Figure 2. This is the case. Note that, separate from the process shown in Figure 3, the scan data output from the distance measuring sensor 11 is also processed. The data is input to the measurement data processing unit 21 at a predetermined interval, and the measurement data processing unit 21, Assume that the input scan data is filtered and then output.

[0025] In step 10 (hereinafter, steps will also be referred to as "S"), the path calculation unit 22 performs the following: Read the environment map 30 from memory 13 and set it.

[0026] In S11, the self-position estimation unit 24 estimates its own position P on the environment map 30. The self-position estimation unit 24 estimates a rough self-position P according to the past travel route. The self-position estimation unit 24 then places the measurement points included in the scan data on the environmental map 30. The position of the object is searched for. At this time, the coordinate transformation unit 23 calculates the position of the object included in the scan data. The coordinates of the measurement point are converted from the robot coordinate system to absolute coordinate system values, and the object on the environment map 30 Compare the location. Then, the position of the object on the environmental map 30 corresponding to the searched measurement point is We use this to estimate our own position P. Note that our own position P is determined by the coordinates (absolute) on the environment map 30. In addition to coordinates, it includes information indicating the attitude of the autonomous robot 100. Self-position estimation unit 24 shows the appearance of the autonomous mobile robot 100, in addition to the above method, as output from the internal sensor 16. Considering information indicating momentum and information indicating the speed of the driving unit 14 detected by the drive control unit 26, With consideration, you may estimate your own position P.

[0027] In S12, the destination is set on the environment map 30. For example, the user displays When the destination is entered by operating 15, the route calculation unit 22 sets the entered destination. In S13, autonomous driving begins. In the autonomous driving initiation performed in S13, the autonomous driving process At the start of the process, the necessary values ​​are set.

[0028] In S14, the state management unit 25 determines whether or not the passing operation has started. Here, Assuming that the passing maneuver has not started (S14: NO), proceed to S17. In S17, the route The calculation unit 22 calculates the travel path that the autonomous mobile robot 100 should follow. In calculating the travel path, by defining a cost map around the device, as described later, On the cost map, the system identifies surrounding obstacles and calculates the travel route and speed.

[0029] Figure 4 is a diagram illustrating the calculation of the travel route in S17. The cost map CM is estimated The arrangement of obstacles is represented by a grid g(m,n) that evenly divides the area around the self-position P. This is the map shown. In this embodiment, the cost map CM is a grid with the current self-position P as the center. Let lid g(0,0) extend by a predetermined length w in the direction x of travel of the autonomous mobile robot 100. It is a rectangular region extending by a predetermined length h in the lateral direction y, which intersects the direction of travel x. In the CM, the position of each grid is determined by dividing the length w in the direction of travel x into equal intervals. It is defined by "m", which represents the distance, and "n", which represents the coordinates obtained by dividing the length h in the horizontal direction y into equal intervals. Yes, for example, the cost map CM extends about 3m forward from its own position P.

[0030] Note that the grid g(m,n) in the cost map CM corresponds to the robot coordinate system. In Figure 4, for the sake of clarity, the cost map CM(t1) calculated at time t1 is shown. ) and the cost map CM(t2) calculated at time t2 are shown, but in reality, In line with the update cycle of the Stormap CM, the Costmap CM is sequentially updated to a position corresponding to its own position P. It will be created in place.

[0031] Specifically, the path calculation unit 22 is determined by the measurement point cloud included in the scan data. The location of the obstacle is converted to the position of grid g(m,n) on the cost map CM. In Figure 4, in each cost map CM(t1) and CM(t2), the areas are filled in black. The grid g(m,n) is the grid where the obstacles are located.

[0032] The path calculation unit 22 uses the scan data to calculate each grid g(m, For n), determine whether there are obstacles. Then, for the searched grid g(m,n) The driving path is calculated by connecting grids g(m,n) that are free of obstacles. The route calculation unit 22 also considers the attitude of its own device and calculates the optimal travel path in the cost map CM. The search is conducted sequentially. In this embodiment, the cost map CM is used to define the location around the self-position P. The defined region is an example of a path calculation region.

[0033] In S18, the route calculation unit 22 calculates the route of the autonomous driving robot 100 based on the route calculated in S17. The drive control unit 26 calculates a command speed to drive according to the route. This information indicates the rotational speeds of the left and right motors 142A and 142B of the drive unit 14. The motion control unit 26 outputs the input command speed to the motor driver 143 of the travel drive unit 14. Drive motors 142A and 142B.

[0034] Furthermore, the position of the autonomous mobile robot 100 also changes according to the control shown in S18, The self-position estimation unit 24 updates its own position P according to the position of the autonomous robot 100 after the change. The self-position estimation unit 24 updates its own position P by detecting the internal sensor 16. The odometry obtained from the results may be used to estimate the self-position P.

[0035] In S19, the route calculation unit 22 determines whether the current position P has reached the destination. The process is terminated. If the current position P has not yet reached the destination (S19:NO), the process returns to S14. The calculation of the travel route in S17 by the route calculation unit 22 and the control in S18 are repeated. The autonomous robot 100 then continues its autonomous journey toward its destination.

[0036] Next, we will explain the process that occurs when a moving object is detected in front of the device during autonomous driving, as shown in Figure 4. The controller 10 performs the process shown in Figure 4 in parallel with the process related to detecting the moving object shown in Figure 5. It is executing.

[0037] In S30, the object detection unit 27 analyzes the imaging data acquired by the imaging unit 12 and captures The specific moving object Ob contained in the data is detected. It is detected by the analysis of the imaging data in S30. The types of specific mobile devices (Ob) are: "healthy individuals," "individuals requiring walking assistance," and "transportation devices accompanied by a transporter." It is one of the following. Figure 6 shows an example of data included in imaging data, where the type is "healthy person". This is a diagram illustrating the specific moving object Ob. "Healthy person" refers to a person included in the imaging data. This is a mobile object that does not qualify as a person requiring walking assistance. "Persons requiring walking assistance" refers to individuals using walking aids such as canes. This refers to a person who is carrying something or a person who is moving in a wheelchair. "Transportation equipment accompanied by a carrier" is a type of transport device. This refers to a mobile entity (Ob) which includes transport equipment such as stretchers and tretchers, and the people who transport them. In the following, when distinguishing between the types of specific mobile bodies (Ob), we will add "Ob1" to the end of the name of a healthy person. Add "Ob2" to the end of the name of the person receiving walking assistance, and add "Ob3" to the end of the name of the transport device that involves a transporter. wear.

[0038] In this embodiment, "walking assistance device" includes canes, walkers, rolling carts, and mobility scooters. The term "cane" includes white canes, T-shaped canes, multi-legged canes, crutches, and Lofstrand crutches. Yes, they are.

[0039] A method for detecting a moving object Ob from imaging data is, for example, from the imaging data shown in Figure 6. When detecting "healthy individuals," a learning model that has been pre-trained on the characteristics of people (healthy individuals) is used. The degree of agreement of the feature quantities of specific moving objects within the imaging data indicates the presence of areas in the imaging data that indicate healthy individuals. Determine whether or not the region is included. On the other hand, using the aforementioned learning model, from the imaging data, If both the area of ​​a person and the area of ​​a cane can be detected, or if both the area of ​​a person and the area of ​​a wheelchair can be detected together... If this occurs, it can be determined that the imaging data includes a "person requiring walking assistance". Furthermore, using the aforementioned learning model, the human area and the stretcher area are identified from the imaging data. If both are detected, the imaging data indicates that "a transport device with a transporter" is included. It is possible to determine this.

[0040] If the specific moving object Ob cannot be detected from the imaging data (S31: NO), return to S30. Continue analyzing the imaging data. Meanwhile, if a specific moving object Ob is detected (S31: YES) ), proceed to S32. In S32, the object detection unit 27 detects a specific moving object from the imaging data. The object detection unit 27 performs tracking processing on Ob. In the tracking processing, the object detection unit 27 uses the imaging data to perform tracking processing. For each specific mobile object (Ob) included, a type ID is assigned to identify its type, and it is also identified as a target for tracking. The object detection unit 27 then assigns a tracking ID that indicates the type ID and the tracking ID. For a specific moving object Ob, the relative distance Ls from the device to this specific moving object and the relative speed Calculate the degree Vs.

[0041] Figure 7 is a diagram illustrating the position of the specific moving object Ob that is the target of the tracking process. 27 performs tracking processing when it detects a specific moving object Ob in front of its own device. In Figure 7, The region A in which tracking is performed for a stationary moving object Ob is defined as the current self-position P (i.e., cost From g(0,0)), which is the center of the tomap CM(tn), the number ahead of this self-position P This is shown as a region defined within the range up to distance L2. The second distance L2 is the imaging unit 12 By imaging the area in front of the device, it becomes possible to detect a specific moving object Ob from the image data. This is a distance that, relative to the self-position P, is forward of the first distance L1 described later. The second distance L2 is, for example, 10 [m]. The first distance L1 is determined in S34, which will be described later. This is a predetermined distance used to determine the distance Ls, and is based on the current self-position P, and the cost This is a position ahead of map CM(tn). The first distance L1 is, for example, 5 [m]. In Figure 7, the length from the self-position P to the tip of the cost map CM(tn) is denoted as L3. It is included. In other words, region A, where tracking processing is initiated for a specific moving object, is currently, This includes locations ahead of the region defined by the cost map CM(tn). In this embodiment, region A is the area around the autonomous mobile robot 100, and the current cost match It is defined to overlap with CM(tn), but it is not limited to this. For example, in region A The end of the autonomous robot 100 should not overlap with the current cost map CM(tn) It may also be stipulated in [the relevant section].

[0042] Returning to Figure 5, in S33, the object detection unit 27 detects the specific moving object Ob that has become the target of tracking. The tracking process is then terminated. Specifically, the object detection unit 27 determines whether to terminate the tracking process based on the image data. If the object detection unit 27 can no longer detect the specific moving object Ob that has become the target of tracking, The process ends (S33:YES), and the process returns to S30. For example, if a specific mobile object Ob is autonomously moving If the autonomous driving robot 100 passes by another robot 100, or changes its course, the autonomous driving robot 10 When the object is separated from 0 by a relative distance L2 or more, the object detection unit 27 terminates the tracking process for the specific moving object Ob. do.

[0043] On the other hand, the object detection unit 27 will not complete the tracking process for the specific moving object Ob until (S33 :NO), proceed to S34. In S34, the object detection unit 27 detects the specific moving object that has been tracked. The relative distance Ls to the autonomous mobile robot 100 relative to Ob has become less than or equal to the first distance L1. The object detection unit 27 determines whether or not it is the target of tracking. If the distance is greater than the first distance L1 (S34: NO), it returns to S32, and the special target of tracking is... Continue tracking the stationary moving object Ob.

[0044] The object detection unit 27 detects that the relative distance Ls of the specific moving object Ob has become less than or equal to the first distance L1. If a decision is made (S34: YES), proceed to S35. The passing motion determination unit 28 then proceeds to S3 The processing from 5 to S41 determines the type of the specific mobile object Ob, and how this specific mobile object Ob approaches. The mode of passing a specific moving object Ob is set according to the movement speed Va at the time of movement.

[0045] First, in S35, the passing motion determination unit 28 determines the type of specific moving object determined in S30. This is a "transportation device accompanied by a transporter," and the movement speed Va of the specific moving object Ob is less than or equal to the first velocity V1. Determine whether or not this is the case. The movement speed Va of the specific mobile object Ob is determined by whether the specific mobile object Ob is autonomously moving. When approaching robot 100, the autonomous mobile robot 100 was set to move in the positive direction. This is the speed in a given case, and the negative side (i.e., the side approaching the autonomous robot 100) This is the speed of the specific mobile object Ob as an autonomous mobile robot 10. Here, "first speed V1" is the speed of the specific mobile object Ob as an autonomous mobile robot 10 As it approaches 0, the autonomous mobile robot 100 will perform a passing maneuver with the specific mobile object Ob. This is the standard value for the expected movement speed when carrying a transport device with a transporter. The system determines whether or not it is approaching at a speed faster than the robotic robot 100.

[0046] Transport equipment accompanied by a person is used when moving at the same speed as a "healthy person" or in emergency transport. It is assumed that they sometimes move at a faster speed than "healthy individuals". Therefore, the first velocity V1 is either the same value as the third speed V3 used in the decision made in S39 described later, or relative to the third speed V3. This results in a faster speed on the negative side. In this embodiment, the first speed V1 is, for example, -3 [km]. It can be set to / h] or more. If the passing motion determination unit 28 determines S35 to be positive, S35:YES), proceed to S36 and set the operation flag indicating the operation mode of passing maneuvers to "Evacuate". Set to "".

[0047] The movement speed Va used for the determination in S35 is the specific moving body detected in the tracking process in S32. It is calculated using Ob's relative velocity Vs and the current movement velocity Vr of the autonomous mobile robot 100. This is possible. As mentioned above, if the direction of movement of the autonomous mobile robot 100 is positive, Movement speed can be calculated as "Va = Vr + Vs". Note that in S37 and S39... The same applies even if they are present.

[0048] On the other hand, if the passing motion determination unit 28 determines that S35 is negative (S35: NO), then S37 Proceeding to the next step, the type of specific mobile body Ob determined in S30 is "person requiring walking assistance," and We determine whether velocity Va is less than 0 and greater than the second velocity V2. That is, walking. Is the person being assisted approaching at a slower speed than the autonomous robot 100? To determine whether or not. In this case, the person receiving walking assistance will move slower than a "healthy person". As this is expected, the second speed V2 is used in the first and third decisions made in S35 and S39 described later. A speed slower than speeds V1 and V3 can be used on the negative side. For example, in S37 The second speed V2 used for the decision can be -1 [km / h]. Passing motion determination. If section 28 makes an affirmative judgment in S37 (S37: YES), it proceeds to S38, and the passing action Set the operation flag indicating the operation mode to "Avoid".

[0049] On the other hand, if the passing motion determination unit 28 determines that S37 is negative (S37: NO), then S39 The process then proceeds, and the type of specific mobile body Ob determined in S30 is "healthy person", and the specific mobile body Ob Determine whether the movement speed Va is less than or equal to the third velocity V3. For example, the third velocity V3 is - It can be set to 3 [km / h]. That is, in S39, healthy person Ob1, autonomous driving robot Determine whether it is approaching at a speed faster than 100. A positive judgment is made in S39 (S39 YES) Then the passing motion determination unit 28 proceeds to S40 and determines the motion of the passing motion. Set the action flag indicating this to "slow down".

[0050] In this embodiment, the object detection unit 27 detects the type of a specific moving object Ob in the order of S35, S37, and S39. By making a distinction, the person being transported, as included in "transportation equipment with a transporter," is classified as "healthy." This can prevent it from being recognized.

[0051] In this embodiment, the object detection unit 27 determines the type of the specific moving object Ob in S30, and then S The operation flag is set based on the judgments in 35, S37, and S39. Alternatively, the object In S30, the detection unit 27 detects a moving object without determining its type from the imaging data, The type of mobile object may be determined in S35, S37, and S39. In this case as well, S35 Then, it is determined whether the type is "a transport device accompanied by a transporter", and in S37, if the type is "walking assistance It is good to determine whether or not the person is an "elephant" and then determine in S39 whether or not their category is "healthy". .

[0052] The passing motion determination unit 28 determines the operation flag by processing in any of S36, S38, or S40. When the setting is selected, the process proceeds to S41, and the start of the passing maneuver is notified. Notifications performed in S41 Next, the value indicated by the operation flag is notified to the state management unit 25. The process in S41 is terminated. Then, return to S32 and continue the follow-up process. Meanwhile, make a negative judgment (S39: NO) in S39. In this case, the passing operation determination unit 28 returns to S32 without notifying the value of the operation flag.

[0053] Returning to Figure 3, when the status management unit 25 obtains the value of the operation flag via notification, the passing movement The decision to start the operation is made (S14: YES), and the process proceeds to S15, according to the type of specific mobile object Ob. The passing maneuver is performed. Figure 8 is a flowchart illustrating the procedure of the process performed in S15. It is.

[0054] In S50, the state management unit 25 determines the value of the notified operation flag. If the passing action indicated by the action flag is "slow down" (S51:YES), S52 The robot proceeds and performs a "slow-down maneuver" as a passing maneuver. This action reduces the vehicle's speed to a slow, crawling speed.

[0055] As shown in Figure 9, in the scenario where the operation flag is set to "slow," healthy person Ob1 This is a scene where the robot approaches at a speed faster than autonomous robot 100. Healthy person Ob1, It can move faster than autonomous robot 100 and can freely change its course. Therefore, rather than the autonomous robot 100 avoiding healthy person Ob1, it deliberately approaches healthy person Ob1. It would be better to have the autonomous mobile robot 100 pass by. By reducing the vehicle's speed to a crawling speed, it encourages able-bodied individuals to change their course.

[0056] The state management unit 25, during the slow-movement operation in S52, uses the object detection unit 27 to process the image data From there, maintain a straight line until you can no longer follow healthy person Ob1, which has been assigned a type ID and follow ID. It is sufficient to continue the slow-moving operation as is. In other words, healthy person Ob1, autonomous driving robot 1 By passing by 00, healthy person Ob1 is positioned outside the imaging range of the imaging unit 12, and object detection is performed. The output unit 27 terminates the tracking process for the healthy person Ob1. The display control unit 29 during slow-movement operation. Display text or an image such as an icon indicating that the vehicle should proceed slowly towards healthy individuals (Ob1). It may also be displayed on Ray 15.

[0057] Figure 10 shows, as a comparative example, the autonomous mobile robot 100, without performing a slow-speed operation, This explains how autonomous driving would occur if a healthy person (Ob1) entered the cost map CM(tn+1). This is the figure. Note that the cost map CM(tn+1) is shown in Figure 9. This is a cost map created after the generation of ). In this example, healthy person Ob1 is cost When entering map CM(tn+1), the path calculation unit 22 determines that the healthy person Ob1 To ensure they pass each other, the travel route is calculated using scan data output from the distance measuring sensor 11. At this point, healthy person Ob1 moves in a direction that allows it to pass by autonomous robot 100. If this is changed, the movement speed of the autonomous mobile robot 100 will remain normal, and the relationship between the two If the distance is close, healthy person Ob1 will not be able to avoid a collision with autonomous robot 100. There is a risk. Also, depending on the method used to calculate the travel route, the route changed by healthy person Ob1 may be affected. In some cases, the autonomous robot 100 may calculate a travel path that leads to the target. To avoid this situation, it may be possible to increase the range of the cost map CM(tn). However, by increasing the cost map CM, the route calculation unit 22 may actually... Calculating the row path becomes more complex, and the processing load may increase.

[0058] In this embodiment, the travel route is calculated using a cost map CM in front of the device. The vehicle detects a specific moving object Ob in area A, which is ahead of the designated area, and begins to move slowly. This buys time for the normal Ob1 to make a significant change in course, increasing the possibility of a collision between the two. It can be lowered.

[0059] Returning to Figure 8, if the passing action indicated by the action flag is "avoidance" and not slowing down, then (S 53:YES), the state management unit 25 proceeds to S54 and performs an avoidance operation as a passing operation. The avoidance maneuver involves the autonomous mobile robot 100 following the travel path calculated in S17, in conjunction with the walking assistance countermeasure. This action changes the travel path so as not to obstruct the path of the elephant Ob2.

[0060] As shown in Figure 11, in situations where the action flag is set to "avoid," the person receiving walking assistance is... In this scene, the subject Ob2 is slowly approaching the autonomous mobile robot 100. In this scenario, the person receiving walking assistance, Ob2, moves at a speed equal to that of the autonomous mobile robot 100. Because it can only move at a speed slower than its actual speed, and has difficulty changing direction, autonomous driving It would be better if robot 100 changed its course so that it could pass by the side of the person Ob who is receiving walking assistance. Therefore, while maintaining the movement speed of the autonomous mobile robot 100, the calculation in S17 is performed. The travel route will be changed to a route that does not obstruct the path of Ob2, the person receiving walking assistance. , the area in front of the device that is ahead of the region where the travel path is calculated using the cost map CM In the area, the autonomous mobile robot 10 detects the person Ob2 who is receiving walking assistance and initiates an avoidance maneuver. This buys time for the vehicle to make a significant change in course, reducing the likelihood of a collision between the two. It is possible.

[0061] In this embodiment, the state management unit 25 avoids the walking assistance recipient Ob2 used in S54. The object detection unit 27 performs a tracking process to determine the travel path, assigning a type ID and a tracking ID to the path. The location of healthy person Ob1 is determined using the calculated travel route by the route calculation unit 22. Furthermore, the state management unit 25 detects the object detection unit 27 from the imaging data to identify the person Ob2 who is receiving walking assistance. If it becomes impossible to display, the avoidance operation in S54 should be terminated. The display control unit 29 controls the avoidance operation Inside, there is text or an eye indicating that the vehicle is traveling along a route that avoids the walking assistance recipient, Ob2. It may also be used to display images such as a computer screen on the display 15.

[0062] Returning to Figure 8, the passing action indicated by the action flag determined in S50 is whether it is a retreat or not. (S53:NO), proceed to S55, and perform a "retreat action" as a passing maneuver. Retreat action The autonomous robot 100 was positioned to the side of the passage so as not to obstruct the path of the specific mobile object Ob. This is an operation where the object remains stationary while in motion until a specific moving object (Ob) passes by.

[0063] As shown in Figure 12, in situations where the operation flag is set to "evacuate", the transporter is present. This scene shows the transport device Ob3 approaching the autonomous mobile robot 100 at high speed. In this scenario, the transport device Ob3, which carries a person, is moving at a faster speed than the autonomous robot 100. It is capable of moving, but its course changes are difficult. Therefore, autonomous mobile robot 100 However, by stopping to the side of the passage so as not to obstruct the path of the transport device Ob3 with the transporter, It would be better to have the autonomous mobile robot 100 pass by the transport device Ob3, which is carrying a person.

[0064] In this embodiment, the state management unit 25 controls the transport device Ob3 with a transporter used in S55. The object detection unit 27 determines a travel path to move away from the path by performing a tracking process that determines the type ID and The position of healthy person Ob1, which is assigned a tracking ID, and the route calculated by the route calculation unit 22 are used. The decision is made based on the following. In addition, the status management unit 25 determines that the object detection unit 27 is accompanied by the transporter based on the imaging data. When the transport device Ob3 can no longer be detected, the retraction operation in S55 should be terminated. In the modified travel path of the autonomous robot 100, there are obstacles (for example, If there are waiting benches or open doors, the autonomous robot 100 will move through the passageway. Alternatively, the vehicle may be moved to the side where the obstacle is located and stopped before reaching the obstacle. Furthermore, the display control unit 29 will instruct pedestrians carrying the transport device Ob3 with a person to move to safety. This involves displaying text or an image such as an icon on the display 15 to indicate what to do. It is also possible to do so. In this case, when the evacuation operation is completed and the system stops, the display 15 The image shows the autonomous mobile robot 100 with its display facing the person transporting the stretcher. It is best to stop while the momentum has changed. By doing so, the stretcher and this stretcher While indicating that pedestrians carrying the stretcher have been evacuated, the stretcher After passing by, it can smoothly return to autonomous driving.

[0065] After executing processes S52, S54, and S55, the process proceeds to S16 in Figure 3. In S16, passing each other occurs. Determine whether the operation has finished or not, and if the passing operation has not finished (S16:NO), Return to S15 and continue passing each other. Meanwhile, once the passing each other operation is complete (S16:Y ES), proceed to S17. As mentioned above, in the processing of S52, S54, and S55, autonomous driving The robot 100 passes by specific mobile objects (healthy person Ob1, person requiring walking assistance Ob 2. When a transport device (Ob3) with a transporter can no longer be detected from the imaging data, The operation is now complete.

[0066] In S17, the path calculation unit 22 calculates the current self position estimated by the self position estimation unit 24. The travel route is calculated using P and scan data. In S18, the route calculation unit 22 The drive control unit 26 is instructed to drive according to the driving path calculated in S17, and the command speed is set accordingly. Calculate. As a result, the autonomous mobile robot 100 will switch from driving by passing other vehicles to S1 The vehicle resumes driving according to the driving path calculated in step 7. The state management unit 25 checks its current position P When it is determined that the destination has been reached (S19: YES), the process proceeds to S20, and the drive control unit 26... , and stop autonomous driving.

[0067] The embodiments described above can achieve the following effects. The autonomous mobile robot 100 detects a specific mobile object Ob approaching it from the front. In this case, the current cost map CM(tn) defined around the specific mobile object Ob This describes the behavioral patterns of a specific mobile object (Ob) that occur before it enters the area, depending on its type. This performs a passing maneuver. This prevents the two objects from approaching each other without passing each other. It is possible.

[0068] A scene in which a specific mobile entity, Ob, which has been determined to be healthy, approaches an autonomous robot. Therefore, by having the autonomous robot 100 move slowly, it can avoid passing healthy people appropriately. It can perform certain actions.

[0069] A specific mobile object Ob, which was determined to be a person requiring walking assistance, approached the autonomous mobile robot 100. In situations where this is the case, the autonomous robot will perform evasive maneuvers to assist the person receiving walking assistance. This allows for the proper passing maneuver to be performed.

[0070] A specific mobile object Ob, identified as a transport device accompanied by a carrier, approached an autonomous mobile robot. In the scene in question, the autonomous mobile robot 100 does not obstruct the path of the transport equipment carrying the transporter. By performing a retreating maneuver, the system ensures proper passing maneuvers for transport equipment that carries a person. It can be executed.

[0071] Only for specific moving objects Ob that are moving in a direction approaching the device, this specific moving object Before entering the area of ​​cost map CM(tn), a passing maneuver is initiated. This prevents unnecessary control measures from being executed to pass by specific moving objects.

[0072] (Modified version of the first embodiment) In the first embodiment described above, as shown in Figure 5, the object detection unit 27 identifies the object in S30. The passing motion determination unit 28 determines the type of moving object Ob, and makes the determination in S35, S37, and S39. The value of the operation flag was set according to the type of specific moving object Ob that had already been processed. In S30, the detection unit 27 does not determine the type of the specific moving object Ob, and the relative distance Ls is the first distance L If it is determined that the value is 1 or less (S34: YES), the type of the specific mobile body Ob is determined to be "carrier Even if you try to determine at once whether someone is using a transport device, a walking assistant, or a healthy person, Good. Then, the state management unit 25 determines the type of the specific mobile object Ob, and the operation flag You can set the setting to one of the following: "Evacuate," "Avoid," or "Slow Down."

[0073] In the first embodiment described above, the passing maneuver is performed taking into consideration the movement speed of the specific moving object Ob. The presence or absence of this was toggled. Instead of this, without considering the movement speed of a specific moving object Ob, You may switch whether or not to perform the different actions. In this case, the person receiving walking assistance, Ob2, The transport device Ob3, which involves a human handler, has less flexibility in changing its course than the autonomous mobile robot 100. Therefore, in Figure 5, if the type of specific mobile object Ob is a transport device accompanied by a transporter, S35:YES), regardless of movement speed Va, proceed to S36 and set the operation flag to save. That is sufficient. Similarly, if the type of specific mobile body Ob is a person requiring walking assistance (S37: YES) Regardless of movement speed Va, you should proceed to S38 and set the action flag to avoid.

[0074] In the first embodiment described above, the mode of passing motion is determined according to the type of specific moving object Ob. Changed. Instead, when a specific moving object Ob is detected, the same type of passing action will be performed. It is permissible to do so. In this case, in S35 of Figure 5, the specific mobile body Ob is a "healthy person", If it is either a person requiring walking assistance or a transport device accompanied by a transporter, set the operation flag to slow down. You can set it to either avoid or retreat. Also, if you omit the processing in S37-S40 good.

[0075] In the tracking process at S32 in Figure 5, if multiple specific moving objects Ob are being tracked, the tracking process Among the specific mobile objects Ob targeted, the first one whose relative distance Ls becomes less than or equal to the first distance L1 For the stationary moving object Ob, the processes S36 to S40 should be executed. In this case, As a passing operation performed by the status management unit 25, first when the relative distance becomes less than or equal to the first distance L1 Depending on the type of specific moving object Ob, a passing maneuver should be performed.

[0076] (Other embodiments) The technologies disclosed herein are not limited to the embodiments described above, and their essence may be omitted. Within the limits of not escaping, it can be transformed into various forms, and for example, the following transformations are also possible: be. In the above embodiment, a hospital was used as an example of an indoor setting. Alternatively, an autonomous mobile robot The indoor areas where the TT100 operates are: "pedestrians," "people requiring walking assistance," and "transportation equipment with a carrier." It can be applied to any indoor location where there is foot traffic, including public buildings. It may be a hotel or an apartment building. In this case, "transportation equipment with a transporter" is transport Transport equipment such as wheeled carts, wheeled cargo carts, folding trolleys, and strollers. It consists of either a large carrier bag or a transporter who transports this transport equipment. It's okay to be there.

[0077] The shape of the autonomous driving robot 100 is not limited to the external shape shown in Figure 1. Furthermore, autonomous driving The robot 100 does not need to be equipped with a display 15.

[0078] In the above-described embodiment, the autonomous mobile robot 100 is equipped with only one controller 10. This was the configuration. However, the autonomous mobile robot 100 is not limited to this configuration and is equipped with multiple controllers. Alternatively, each controller may be responsible for executing different functions. [Explanation of Symbols]

[0079] 10...Controller, 11...Distance sensor, 12...Imaging unit, 13...Memory, 14...Driving drive Moving part, 15...Display, 16...Internal sensor, 21...Measurement data processing unit, 22...Path meter Calculation unit, 23... Coordinate transformation unit, 24... Self-position estimation unit, 25... State management unit, 26... Drive control unit ,27...Object detection unit, 28...Passing motion determination unit, 100...Autonomous mobile robot, CM...Co Streetmap

Claims

1. A sensor that detects objects located around the device, An imaging unit that acquires imaging data of the area in front of the device, The drive unit and An autonomous driving control unit calculates a driving path in a path calculation area around the device based on the position of an object detected by the aforementioned sensor, and drives the driving drive unit according to the calculated driving path, thereby causing the device to drive autonomously. During autonomous driving, the device includes an object detection unit that detects a specific moving object located ahead of the path calculation area relative to the device itself, based on the image data, and determines the type of the detected specific moving object. The object detection unit, in order, identifies the specific moving objects included in the imaging data as: a transport device with a carrier, a person holding a walking assistance device or sitting in a wheelchair, a person receiving walking assistance, and a healthy person who is not holding a walking assistance device and is not sitting in a wheelchair. The autonomous driving control unit controls the driving drive unit to cause the autonomous driving robot to perform a passing maneuver in which it passes the detected specific mobile object in an operating manner corresponding to the type of specific mobile object that has been determined.

2. The autonomous driving robot according to claim 1, wherein when the object detection unit determines that the specific moving object is a transport device accompanied by a transporter, the autonomous driving control unit controls the driving drive unit to perform a retraction operation as the passing operation, such as retracting the object to avoid obstructing the path of the transport device accompanied by the transporter.

3. The object detection unit identifies a person holding a walking assistance device or a person in a wheelchair as a specific moving object included in the imaging data, and identifies them as a person who needs walking assistance. The autonomous driving robot according to claim 1, wherein when the autonomous driving control unit determines that the specific moving object is the person to be assisted in walking, the autonomous driving control unit controls the driving drive unit to perform an avoidance operation as the passing operation, such as driving along a driving path that avoids the path of the person to be assisted in walking.

4. The object detection unit identifies a person who is not holding a walking aid and is not in a wheelchair as a specific moving object included in the imaging data, and identifies them as a healthy person. The autonomous driving control unit controls the driving drive unit to perform a slow-movement operation, which reduces the speed of the device to a slow-movement speed as a passing operation, when the object detection unit determines that the specific moving object is a healthy person, according to claim 1.

5. The autonomous driving control unit, An autonomous mobile robot according to any one of claims 1 to 4, wherein the mobile drive unit is controlled to perform the passing maneuver when the specified mobile body is moving in a direction approaching the device.

6. A sensor that detects objects located around the device, An imaging unit that acquires imaging data of the area in front of the device, A program executed by the controller of an autonomous mobile robot equipped with a driving unit, The aforementioned controller, Based on the position of the object detected by the aforementioned sensor, the autonomous driving control process calculates a driving path in the path calculation area around the device, and drives the driving drive unit according to the calculated driving path, thereby causing the device to drive autonomously. During autonomous driving, an object detection process is performed to detect a specific moving object located ahead of the path calculation area relative to the device, based on the image data, and to determine the type of the detected specific moving object. Make it run, In the object detection process described above, the specific moving objects included in the imaging data are sequentially identified as follows: a transport device with a carrier, a person holding a walking assistance device or sitting in a wheelchair, and a person not holding a walking assistance device and not sitting in a wheelchair, as a person requiring walking assistance; and a person not holding a walking assistance device and not sitting in a wheelchair, as a healthy person. The autonomous driving control process includes a program that controls the driving drive unit in order to cause the device to perform a passing maneuver in which it passes the detected specific moving object in an operating manner corresponding to the type of the identified specific moving object.

Images