Method for controlling autonomous mobile robot, and autonomous mobile robot
By receiving near-field obstacle information to generate an obstacle location map and searching for accessible directions, autonomous mobile robots can quickly get out of trouble in complex areas, solving the problem of low elimination efficiency in the prior art and improving work efficiency.
Patent Information
- Application Number
- PCT/CN2024/091983
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-09-16
- Filing Date
- 2024-05-09
- Publication Date
- 2024-11-14
AI Technical Summary
The autonomous mobile robot has low efficiency in getting out of trouble in complex areas, and the existing technology's backward, rotation and forward strategies are poorly targeted, resulting in low efficiency in getting out of trouble in complex areas and affecting the user experience.
By receiving near-field obstacle information sent by the sensor, a map of obstacle location is generated, and based on the current location and map search for the passable direction, the expansion coefficient and search range are used to adjust the path to ensure that the robot moves in the passable direction until it gets out of trouble.
It improves the efficiency of autonomous mobile robots to escape difficulties in complex areas, avoids inefficient escape caused by random impacts, and ensures that the robot can quickly and targetedly leave the difficulties.
Smart Images

Figure CN2024091983_14112024_PF_FP_ABST
Abstract
Description
Control method of autonomous mobile robot, autonomous mobile robot Technical Field
[0001] The invention relates to a control method of an autonomous mobile robot and the autonomous mobile robot. Background Art
[0002] Autonomous mobile robots are capable of autonomously moving within a defined area, performing tasks while on the move. Because they operate without human intervention, they may enter complex areas during their movements, becoming trapped and unable to move further. In such situations, autonomous mobile robots typically employ a strategy of retreating, rotating, and then moving forward to escape. However, this method of escape is less targeted, resulting in low efficiency in complex areas and a negative user experience.
[0003] Summary of the Invention
[0004] In order to overcome the defects of the prior art, the problem to be solved by the present invention is to provide an automatic working method and an autonomous mobile robot for efficiently escaping from trouble.
[0005] The technical solution adopted by the present invention to solve the problems of the prior art is: a control method for an autonomous mobile robot, characterized in that it includes: receiving near-field obstacle information sent by at least one sensor, and the at least one sensor is installed on the autonomous mobile robot; generating an obstacle position map around the autonomous mobile robot based on the near-field obstacle information; generating an escape path in response to receiving the near-field obstacle information multiple times within a limited time or space, wherein the generating of the escape path includes: searching for a passable direction based on the current position of the autonomous mobile robot and the obstacle position map.
[0006] In one embodiment, the near-field obstacle information includes at least one of the following: collision information representing a contact collision of the autonomous mobile robot, lifting information representing the autonomous mobile robot being lifted, motor torque information representing a stall of the autonomous mobile robot's driving motor, and obstacle remote sensing information representing the existence of obstacles around the autonomous mobile robot.
[0007] In one embodiment, generating an obstacle location map around the autonomous mobile robot based on the near-field obstacle information includes: marking locations on the map corresponding to the near-field obstacle information as near-field obstacles.
[0008] In one embodiment, the obstacle map is a local map of the periphery of the autonomous mobile robot, and the local map is updated as the autonomous mobile robot moves.
[0009] In one embodiment, based on the current position of the autonomous mobile robot and the obstacle location map, a traversable direction is searched, including: expanding the near-field obstacles in the obstacle location map according to an expansion coefficient, wherein the expansion coefficient is approximately equal to the width of the autonomous mobile robot; and using the area in the obstacle location map where the near-field obstacles do not exist as the traversable direction.
[0010] In one embodiment, expanding the near-field obstacles in the obstacle location map according to the expansion coefficient includes: expanding part of the near-field obstacles in the obstacle location map according to the expansion coefficient, wherein the part of the near-field obstacles includes obstacle remote sensing information representing obstacles existing around the autonomous mobile robot.
[0011] In one embodiment, based on the current position of the autonomous mobile robot and the obstacle location map, a passable direction is searched, including: obtaining the search direction and search range for near-field obstacles in the obstacle location map, and taking the direction in which there are no near-field obstacles within the search range as the passable direction, the initial search direction is the direction opposite to the moving direction of the autonomous mobile robot, and the search range is the area determined by the search direction and beyond the area covered by the autonomous mobile robot on the map.
[0012] In one embodiment, when the passable direction cannot be found in the obstacle location map along the initial search direction, the search direction is switched to a direction opposite to the initial search direction.
[0013] In one embodiment, the method further includes: keeping the search direction unchanged until the autonomous mobile robot receives the near-field obstacle information again while traveling in the passable direction.
[0014] In one embodiment, based on the current position of the autonomous mobile robot and the obstacle position map, a passable direction is searched, including: in the process of controlling the autonomous mobile robot to move in the passable direction, when the near-field obstacle information is received within a duration period, the search direction is switched to a direction opposite to the initial search direction.
[0015] In one embodiment, searching for a traversable direction based on the current position of the autonomous mobile robot and the obstacle position map includes adjusting the traversable direction accordingly as the current position of the autonomous mobile robot and the obstacle position map are updated.
[0016] In one embodiment, based on the current position of the autonomous mobile robot and the obstacle position map, a passable direction is searched, including: when the autonomous mobile robot cannot search the passable direction, controlling the autonomous mobile robot to ignore the obstacle position map and adjust the posture, the posture adjustment method including at least one of the following: forward, backward, and rotation; searching for a passable direction according to the current position of the autonomous mobile robot and the obstacle position map after the posture is adjusted.
[0017] In one embodiment, a traversable direction is searched based on the current position of the autonomous mobile robot after adjusting its posture and the obstacle position map, comprising: searching for a traversable direction based on the current position of the autonomous mobile robot after adjusting its posture and the updated obstacle position map; wherein the obstacle position map is obtained by updating the obstacles in the obstacle position map according to a preset rule, and the preset rule is that when a near-field obstacle identified in real time by an obstacle remote sensing sensor installed on the autonomous mobile robot is inconsistent with an obstacle in the obstacle position map, the obstacle position map is updated according to the real-time identification result of the obstacle remote sensing sensor.
[0018] In one embodiment, the moving angular velocity and the moving linear velocity are determined based on the angular difference between the moving direction of the autonomous mobile robot and the passable direction, and the autonomous mobile robot is controlled to move toward the passable direction according to the moving angular velocity and the moving linear velocity, wherein the moving angular velocity is positively correlated with the angular difference, and the moving linear velocity is negatively correlated with the angular difference.
[0019] In one embodiment, a passable direction is searched based on the current position of the autonomous mobile robot and the obstacle position map, including: when controlling the autonomous mobile robot to move toward the passable direction, determining whether the autonomous mobile robot is rotatable based on the obstacle position map around the autonomous mobile robot; if it is rotatable, controlling its rotation, and searching for the passable direction based on the position after rotation.
[0020] In one embodiment, when determining whether it is rotatable, the autonomous mobile robot initially determines the rotation direction in accordance with the edgewise direction of the autonomous mobile robot.
[0021] In one embodiment, searching for the traversable direction according to the rotated position includes:
[0022] During the process of controlling the rotation of the autonomous mobile robot, determining whether there is a near-field obstacle ahead by using obstacle remote sensing information collected in real time by an obstacle remote sensing sensor, wherein the obstacle remote sensing sensor is mounted on the autonomous mobile robot and is configured to identify obstacle remote sensing information surrounding the autonomous mobile robot;
[0023] When it is determined that the near-field obstacle does not exist, the direction is used as the passable direction, and the autonomous mobile robot is controlled to move along the direction.
[0024] In one embodiment, it is determined whether the near-field obstacle exists ahead by using obstacle remote sensing information collected in real time by an obstacle remote sensing sensor and the obstacle location map.
[0025] In one embodiment, the method further includes: when the autonomous mobile robot travels at least a first preset distance along the escape path, determining that the autonomous mobile robot has successfully escaped.
[0026] In one embodiment, the method further includes: after the autonomous mobile robot successfully escapes, when it receives the near-field obstacle information multiple times within a limited time or space, marking the area covered in the obstacle map as an obstacle, and planning the path according to the re-marked map.
[0027] Another technical solution adopted by the present invention to solve the problems of the prior art is: an autonomous mobile robot, characterized in that the autonomous mobile robot includes: at least one sensor, the sensor is configured to obtain and send near-field obstacle information, and the at least one sensor is installed on the autonomous mobile robot; a controller, the controller is connected to the sensor signal, and is used to receive the near-field obstacle information, and generate an obstacle location map around the autonomous mobile robot based on the near-field obstacle information; in response to receiving the near-field obstacle information multiple times within a limited time or space, an escape path is generated, wherein the generating of the escape path includes: searching for a passable direction based on the current position of the autonomous mobile robot and the obstacle location map.
[0028] Another technical solution adopted by the present invention to solve the existing technical problems is: a computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by the processor to implement the control method of the autonomous mobile robot described above.
[0029] Compared to existing technologies, the present invention has the following advantages: receiving near-field obstacle information transmitted by at least one sensor, generating an obstacle location map around the autonomous mobile robot based on the near-field obstacle information; and generating an escape path in response to receiving near-field obstacle information multiple times within a limited time or space, wherein generating the escape path includes searching for a navigable direction based on the autonomous mobile robot's current position and the obstacle location map. Compared to methods that detect a robot in distress and control it to randomly swerve in various directions to escape, the method of the present invention allows the lawn mower to escape quickly, improving work efficiency.
[0030] The present application also provides a control method for a self-moving device, a self-moving device and a readable storage medium, which detect whether there are obstacles in the area to be driven in front of the self-moving device, and then control the self-moving device to drive out of the area without obstacles to escape. Due to visual recognition, the driving of the self-moving device when getting out of trouble is more targeted, which can effectively improve the escape efficiency of the self-moving device in complex areas.
[0031] To solve the above technical problems, the present application provides a control method for a self-moving device, wherein the self-moving device includes a visual sensor, and the visual sensor is used to obtain an image in front of the self-moving device. The control method includes:
[0032] Detecting whether the self-moving device meets the conditions for escaping;
[0033] If the conditions for escaping are met,
[0034] Detecting whether there is an obstacle within a range corresponding to the area to be driven in the image in front of the mobile device;
[0035] If there is no obstacle within the range corresponding to the area to be driven in the image, the self-moving device is controlled to drive out of the area to be driven to escape.
[0036] Optionally, the method further includes:
[0037] If there is an obstacle within the range corresponding to the area to be driven in the image, controlling the self-moving device to rotate;
[0038] The step of detecting whether there is an obstacle within a range corresponding to the area to be driven in the image in front of the mobile device is started.
[0039] Optionally, if an obstacle exists within a range corresponding to the area to be driven in the image, controlling the self-moving device to rotate includes:
[0040] If an obstacle exists within the range corresponding to the area to be driven in the image, determining whether the self-moving device has rotated by a first preset angle in the current rotation direction;
[0041] If the first preset angle is not rotated, the self-moving device is controlled to rotate according to the current rotation direction.
[0042] Optionally, if an obstacle exists within a range corresponding to the area to be driven in the image, controlling the self-moving device to rotate further includes:
[0043] If the device has rotated by the first preset angle, the area to be driven is reduced, and the self-moving device is controlled to rotate in the opposite direction of the current rotation direction.
[0044] Optionally, if there is no obstacle within the range corresponding to the area to be driven in the image, after controlling the self-moving device to drive out of the area to be driven to escape, the method further includes:
[0045] Determining whether the distance traveled by the self-moving device from the area to be traveled is less than a first preset distance;
[0046] If the distance is less than the first preset distance, it is determined that the self-moving device meets the conditions for escaping, and the number of escape attempts is accumulated;
[0047] The method further comprises:
[0048] Before controlling the self-moving device to rotate, determining whether the number of escape attempts corresponding to the current rotation direction has reached a preset number;
[0049] If the preset number of times is reached, the self-mobile device is controlled to rotate and the image acquisition is paused;
[0050] When the self-mobile device rotates to a second preset angle, image acquisition is restarted and the number of escape attempts is reset to zero.
[0051] Optionally, the method further includes at least one of the following:
[0052] If the self-moving device is currently in a scene where it is trapped by dense grass, before the self-moving device is controlled to rotate for the first time, the cutter head of the self-moving device is controlled to stop and retreat, and when the self-moving device retreats a preset distance, the cutter head of the self-moving device is controlled to start and rotate;
[0053] If the self-moving device is currently in a slipping and trapped scenario, the self-moving device is controlled to rotate at a first speed for the first time, and a different speed is used in each rotation, where the first speed is less than a preset speed.
[0054] Optionally, the method further includes:
[0055] If there is an obstacle within the range corresponding to the area to be driven in the image, reducing the area to be driven;
[0056] Enter the step of detecting whether there is an obstacle within the range corresponding to the area to be driven in the image.
[0057] Optionally, one side of the area to be driven before reduction is larger than the body width of the self-moving device, and the other side is a second preset distance; one side of the area to be driven after reduction is smaller than the body by a preset width or equal to the body width, and the other side is a third preset distance, and the third preset distance is smaller than the second preset distance.
[0058] Optionally, if there is no obstacle within the range corresponding to the area to be traveled in the image, controlling the self-moving device to travel out of the area to be traveled to escape, includes:
[0059] If there are no obstacles within the range corresponding to the area to be driven in the image and the vehicle is currently in a preset trapped scene, determining whether the rotation angle in the current rotation direction is greater than a third preset angle;
[0060] If the rotation angle is greater than a third preset angle, the self-moving device is controlled to drive out of the waiting driving area to escape.
[0061] Optionally, the method further includes:
[0062] If the rotation angle is less than or equal to the third preset angle, controlling the self-moving device to rotate according to the current rotation direction;
[0063] The step of detecting whether there is an obstacle within a range corresponding to the area to be driven in the image in front of the mobile device is started.
[0064] Optionally, after detecting whether the self-moving device meets the conditions for escaping, the method further includes:
[0065] If the conditions for escaping are met, the trapped scene of the self-moving device is determined based on the conditions for escaping that the self-moving device meets; the conditions for escaping include at least one of the following:
[0066] The self-moving device experiences collisions greater than a preset number of times;
[0067] The driving wheel of the self-moving device is blocked;
[0068] A lifting sensor detects that the self-moving device is lifted;
[0069] When the self-propelled device is mowing in situ, the actual rotation speed of the cutter disc is less than or equal to the preset rotation speed;
[0070] The image acquired by the visual sensor remains unchanged for a preset time period.
[0071] Optionally, after controlling the self-moving device to drive out of the waiting area to escape, the method further includes:
[0072] Determine whether the escape was successful;
[0073] If the escape is successful, the self-moving device is controlled to start working.
[0074] The present application also provides a self-moving device, the self-moving device comprising a visual sensor and a controller;
[0075] The visual sensor is used to obtain an image in front of the mobile device;
[0076] The controller is used to implement the steps of any of the above-mentioned methods for controlling a self-moving device when executed.
[0077] The present application also provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of any of the above-described methods for controlling a self-mobile device are implemented.
[0078] The present application discloses a control method for a self-moving device, a self-moving device, and a readable storage medium. The self-moving device includes a visual sensor, which is used to obtain an image in front of the self-moving device. The control method includes: detecting whether the self-moving device meets the conditions for getting out of trouble; if the conditions for getting out of trouble are met, detecting whether there are obstacles in the range corresponding to the area to be driven in the image in front of the self-moving device; if there are no obstacles in the range corresponding to the area to be driven in the image, controlling the self-moving device to drive out of the area to be driven to get out of trouble. The present application detects whether there are obstacles in the area to be driven in front of the self-moving device based on the image obtained by the visual sensor, and then controls the self-moving device to drive out of the area without obstacles to get out of trouble. Since visual recognition makes the driving direction of the self-moving device more targeted when getting out of trouble, it can effectively improve the efficiency of the self-moving device in getting out of trouble in complex areas. BRIEF DESCRIPTION OF THE DRAWINGS
[0079] The above-mentioned objectives, technical solutions and beneficial effects of the present invention can be achieved through the following drawings:
[0080] FIG1 is a schematic diagram of an application scenario according to an embodiment of the present invention;
[0081] FIG2 is a schematic diagram of an automatic working method according to an embodiment of the present invention;
[0082] FIG3 is a schematic diagram of an obstacle marking according to an embodiment of the present invention;
[0083] FIG4 is a schematic diagram of a robot in a difficult situation according to an embodiment of the present invention;
[0084] 5-6 are schematic diagrams of sectors according to an embodiment of the present invention;
[0085] FIG7 is a schematic diagram of a scenario in which a robot escapes from a difficult situation according to an embodiment of the present invention;
[0086] FIG8 is a schematic flow chart of a control method for an autonomous mobile robot according to an embodiment;
[0087] FIG9 is a schematic diagram showing a detection scenario according to an embodiment. DETAILED DESCRIPTION
[0088] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0089] The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. In addition, the terms "and / or" used herein include any and all combinations of one or more related listed items.
[0090] As shown in Figure 1, the automatic operation system of this embodiment may include: an autonomous mobile robot 1, a boundary 7, and a base station 3. The autonomous mobile robot 1 moves and operates within a working area 4 defined by the boundary 7. The base station 3 can be used to return to replenish energy when the autonomous mobile robot is running low on energy. The boundary 7 may be the perimeter of the entire working area, referred to as the outer boundary. It is typically connected end to end, enclosing the working area 4 and may be electronic or physical. A physical boundary can simply be a natural physical boundary formed by the intersection of the working area 4 and a non-working area, such as the boundary between grass and non-grass, or a boundary formed by a wall, fence, or railing. For example, if a user's lawn is adjacent to a neighbor's, a fence 2 as shown in Figure 1 can be used to form a boundary. An electronic boundary can be established by laying wires around the working area 4 and using a boundary signal generator connected to the wires to generate a virtual boundary signal, such as an electromagnetic signal, an acoustic signal, or a light signal. The working area may also contain areas 5 that are not suitable for the autonomous mobile robot 1 to operate. These areas, such as flower beds, pools, obstacles, etc., can be used to form a boundary. This area 5 is referred to as the inner boundary, and the area outside the inner boundary is the working area.
[0091] The autonomous mobile robot 1 can be an automatic lawn mower, a sweeping robot, an automatic snow blower, or other unattended equipment that automatically moves on the surface of a work area to mow the grass, vacuum, or sweep snow. Of course, the autonomous mobile robot is not limited to an automatic lawn mower, a sweeping robot, or an automatic snow blower, and can also be other unattended equipment, which is not limited in this application.
[0092] In the following specific embodiments, a detailed description is given by taking the automatic working system as an automatic lawn mower system, that is, the autonomous mobile robot 1 as an automatic lawn mower.
[0093] The automatic lawn mower includes a housing, a mobile module, an operating module, a controller, and an energy module. The controller connects to and controls the mobile module and the operating module to enable the automatic lawn mower to move and operate automatically. In an embodiment of the present application, the automatic lawn mower may also include a visual sensor 11 for capturing images. After capturing the images, the automatic lawn mower performs image segmentation to identify grass, non-grass, soil, fallen leaves, fences, stone slabs, etc., and controls the lawn mower to move within the segmented grass or other categories (e.g., stone slabs) that the user has allowed the lawn mower to move. The lawn mower's operating modes may include, but are not limited to, at least one of the following: planned path mode, random path mode, and edge path mode. The lawn mower can use random path mode to mow within the working area. In this mode, the lawn mower moves forward until the visual sensor identifies non-grass. Once non-grass is identified, it is controlled to turn. After completing internal mowing, the lawn mower can also use edge movement mode to control the lawn mower to move along the boundary to cut the grass on the boundary or return to recharge. The automatic lawn mower may also include a position detection sensor (such as RTK, IMU, ODO, etc.) for collecting the current geographic location. The position detection sensor is used to create a map of the working area, and the movement and operation of the machine are controlled according to the map. Of course, the solution in the following embodiment of the present application is also applicable to the case where the automatic lawn mower does not have a visual sensor, and only a position detection sensor is provided on the lawn mower to create a map of the working area.
[0094] As shown in Figure 1, an autonomous lawn mower, while moving within its work area, enters a small area 6, roughly at an inner right angle, bounded by obstacles 8 and 9. When the autonomous lawn mower encounters obstacle 8, it is controlled to retreat according to its obstacle avoidance strategy to avoid the obstacle. During its retreat, the autonomous lawn mower encounters obstacle 9 again, and is controlled to move forward according to its obstacle avoidance strategy to avoid the obstacle. During its forward movement, the autonomous lawn mower encounters obstacle 8 again, and is controlled to move forward according to its obstacle avoidance strategy to avoid the obstacle. This results in the autonomous lawn mower being trapped within this small area 6. In this scenario, the autonomous lawn mower's inability to accurately identify the obstacle-free area leads to multiple collisions with this small area 6, but it is unable to escape. Alternatively, the autonomous lawn mower may intermittently collide with the boundary of its work area due to its inability to accurately identify the boundary. If the boundary is an obstacle such as a wall, the autonomous lawn mower may suffer some damage, but it is unable to escape this predicament. Alternatively, when the automatic lawn mower reaches fence 2 as shown in Figure 1, it may become stuck there for a long time and unable to escape. Alternatively, when the automatic lawn mower reaches a narrow area (such as an area surrounded by table legs), it may collide with an obstacle (such as a table leg) while searching for an exit, unable to escape the predicament. These predicaments may cause some degree of damage to the machine, resulting in lower machine efficiency and over-cutting of some lawns.
[0095] In response to the above problems, the present application proposes a control method for an autonomous mobile robot, in which near-field obstacle information is received from at least one sensor, and at least one sensor is installed on the autonomous mobile robot; based on the near-field obstacle information, an obstacle location map around the autonomous mobile robot is generated; in response to receiving near-field obstacle information multiple times within a limited time or space, an escape path is generated, wherein generating the escape path includes: searching for a passable direction based on the current position of the autonomous mobile robot and the obstacle location map. Through this method, the trapped situation in the above scenario can be solved and the robot can escape from the predicament. Specifically, as shown in Figure 2, the automatic working method includes the following steps:
[0096] S201: Receive near-field obstacle information sent by at least one sensor, where the at least one sensor is installed on an autonomous mobile robot.
[0097] In one embodiment of the present application, an autonomous mobile robot can receive near-field obstacle information from at least one sensor during movement. The sensor can be a real sensor or a virtual sensor that uses an algorithm to identify obstacles, or a sensor that uses an algorithm to collect certain data and receives this data through a controller to identify whether an obstacle has been encountered. In the drawings of this application, the sensor is illustrated using a visual sensor 11 as an example.
[0098] Specifically, near-field obstacle information (hereinafter also referred to as obstacles or obstacles) represents obstacle information that is close to the robot and located around the robot, and may include at least one of the following: collision information representing a contact collision of the autonomous mobile robot, lifting information representing that the autonomous mobile robot is lifted, motor torque information representing that the driving motor of the autonomous mobile robot is blocked, and obstacle remote sensing information representing that there are obstacles around the autonomous mobile robot. The obstacle remote sensing information is obtained through an obstacle remote sensing sensor, and the obstacle remote sensing sensor may include but is not limited to at least one of the following: a visual sensor, an ultrasonic sensor, a radar, etc. Specifically, in the embodiment of the present application, the obstacle remote sensing sensor is a visual sensor, and the obstacle remote sensing information is a picture as an example for explanation. These near-field obstacle information corresponding to the machine may be: detection of a collision with the camera, detection of a drive wheel being blocked, and detection of the universal wheel being lifted by the lifting sensor provided on the universal wheel.
[0099] S202: Generate an obstacle location map around the autonomous mobile robot based on the near-field obstacle information.
[0100] In one embodiment of the present application, the autonomous mobile robot can generate an obstacle location map around the autonomous mobile robot based on near-field obstacle information. Thus, the autonomous mobile robot can subsequently determine an escape path based on the obstacle location map to help the lawn mower quickly escape from the predicament.
[0101] The obstacle location map in this application can be a work area map sent by the user, or it can be a map created by the automatic lawn mower through a satellite positioning sensor or a visual sensor, and the map is marked with obstacle information, which is obtained through the above-mentioned collision information, lifting information, motor torque information or obstacle remote sensing information. The map can be a local map, which can characterize the environment within a certain range centered on the automatic lawn mower (i.e., the area surrounding the automatic lawn mower), and can be updated as the position of the automatic lawn mower changes. The local map can be obtained based on multiple frames of images in the recent period of time, and the local map can be updated at regular intervals, that is, periodically updated. By establishing a local map, machine memory can be saved.
[0102] S203: In response to receiving near-field obstacle information multiple times within a limited time or space, generating an escape path, wherein generating the escape path includes: searching for a passable direction based on the current position of the autonomous mobile robot and the obstacle position map.
[0103] In one embodiment of the present application, in response to receiving near-field obstacle information multiple times within a limited time or space, generating an escape path may include: when it is detected that the automatic lawn mower is trapped, obtaining a map of the area around the lawn mower with obstacles marked on it, and controlling the lawn mower to move to an area in the map that is not marked as an obstacle to escape the predicament.
[0104] In one embodiment of the present application, when a robot receives near-field obstacle information multiple times within a limited time or space, it can be determined that the robot is in trouble. In this case, a traversable direction can be searched based on the current position of the autonomous mobile robot and a map of obstacle locations. After obtaining the traversable direction, a path to the traversable direction can be planned based on the robot's current position to escape the trouble, i.e., escape the trouble. In the case of an inaccurate or incomplete map, searching for a traversable direction can control the robot to escape the trouble more quickly.
[0105] In one embodiment of the present application, the robot can be determined to be in a predicament when it detects that it receives near-field obstacle information multiple times within a preset time period or receives near-field obstacle information continuously for a period of time while moving, or when the robot determines that its surroundings remain substantially consistent within a preset time period based on multiple frames of imagery detected by its visual sensor. Alternatively, if the robot turns left and its IMU detects a change in yaw angle greater than a preset value, such as 540 degrees, and the longest side of the circumscribed rectangle of the movement path where the yaw angle changes is less than or equal to a preset length, such as 5 meters, the robot can be determined to be in a predicament due to being unable to escape. If the robot is in a predicament due to being unable to escape, a collision, a stall, or a lift is detected, the robot can be determined to be in a predicament. When the robot receives the aforementioned types of near-field obstacle information, it can be determined that the robot is in a predicament, and the solution proposed in this application can be used to control the robot to escape.
[0106] In the embodiments of the present application, scenarios in which an autonomous lawn mower finds itself in distress often occur in confined spaces, such as small areas. Whether a collision, lift, or stall is detected that meets duration requirements, or whether a collision, lift, or stall is detected that meets both number and time interval requirements, determines whether the robot is trapped and triggers an escape action. Specifically, if three consecutive collisions occur with the interval between each collision less than 10 seconds, or if collisions are detected for a continuous period of 3 seconds, the robot is considered trapped. If 20 collisions are detected within 5 minutes, the robot is deemed to be in a complex situation and may not be able to handle it on its own. Therefore, it is controlled to shut down and an alarm is issued to alert the user for assistance. Furthermore, to prevent the robot from falsely detecting collisions, a collision is only considered to have occurred if a collision is detected for less than 0.1 seconds. If a lift is detected for less than 10 seconds for three consecutive times, or if a lift is detected for less than 3 seconds, the robot is considered to be trapped. Furthermore, to prevent false lift detections, a lift is only considered to have occurred if a lift is detected for less than 0.2 seconds. When three consecutive stalls occur and the interval between each stall detection is less than 10s, or when stalls are detected for 3 consecutive seconds, it can be considered that it is in trouble (stalls include left wheel stalls or right wheel stalls). Furthermore, in order to avoid the automatic lawn mower from mistakenly detecting a stall, it will only be determined that a stall has occurred when a stall is detected for 3 consecutive seconds. Of course, the detected fault type can also be a combination of collision, lifting or stalling. For example, when the automatic lawn mower detects 1 collision, 1 lifting, and 1 stall in succession, it can be considered that it is in trouble. Of course, the duration and number of times used in the above embodiment to determine that it is in trouble are only examples. According to the improvement of the detection accuracy of the automatic lawn mower, adaptive adjustments can be made to improve its reaction time for starting the escape strategy. For example, the automatic lawn mower can be adjusted from detecting a collision for 3 consecutive seconds to detecting a collision for 2 consecutive seconds. This application does not limit this.
[0107] Furthermore, when it is detected that the automatic lawn mower is trapped, a local map within a certain range centered on it can be obtained. For example, the map can be a rectangular area with a side length of 2m centered on the position of the automatic lawn mower. The position on the map corresponding to the near-field obstacle information received by the automatic lawn mower can be marked as a near-field obstacle. Specifically, the rectangular area can be divided into multiple grids, and the value of each grid can be updated according to whether the automatic lawn mower sensor recognizes an obstacle and the type of obstacle recognized. Specifically, the grid can be assigned a value based on whether there is an obstacle at the actual working area position corresponding to each grid and the type of obstacle. For example, if there is an obstacle, the grid is assigned a value of 1, and if not, it is assigned a value of 0. When the lift information of the robot being lifted is received, the grid is assigned a value of 1 and the grid type is recorded as lifted. When image information of the existence of obstacles around the robot is received, the grid is assigned a value of 1 and the grid type is recorded as the obstacle recognized by the visual sensor.
[0108] Specifically, as shown in FIG3 , when the lawn mower receives at least one of the following near-field obstacle information while moving, the location corresponding to the near-field obstacle information on the map can be marked as a near-field obstacle to generate an obstacle location map around the lawn mower. When the map is a grid map, a fixed number of grids on the map corresponding to the location of the sensor or drive wheel that received the obstacle information can be marked as obstacles, or a corresponding number of grids can be marked as obstacles based on the size of the obstacle or the torque, with a larger number of grids being marked as obstacles. As shown in FIG3 , when near-field obstacle information is received indicating a camera collision, the six grids K in front of the camera's location on the map can be marked as obstacles, and the type of these grids can be marked as collision. When near-field obstacle information is received indicating a universal wheel lift, the six grids in front of the universal wheel's location on the map can be marked as obstacles, and the type of these grids can be marked as lift. When receiving motor torque information indicating a stalled drive wheel motor (i.e., when the motor torque exceeds a preset threshold), the system obtains the mower's speed, direction, and first area covered on the map. Simulates the mower's movement at that speed and direction for a preset time, covering a second area on the map. The system determines the area within the second area that is not covered by the first area, and marks the uncovered area as a near-field obstacle. This area can be a grid. If the uncovered area is half a grid, the entire grid is marked as a near-field obstacle for safety reasons.
[0109] In one embodiment of the present application, searching for a traversable direction based on the autonomous mobile robot's current location and an obstacle location map may include: expanding near-field obstacles in the obstacle location map according to an expansion factor, the expansion factor being approximately equal to the width of the autonomous mobile robot; and determining areas in the obstacle location map where no near-field obstacles exist as traversable directions. By expanding a grid marked as an obstacle, the grid covered by an autonomous lawn mower when it approaches the obstacle is simulated, so that the uncovered area becomes a traversable area, and the lawn mower can be controlled to move in the direction corresponding to that area.
[0110] Furthermore, when expanding, only some of the near-field obstacles in the obstacle location map can be expanded. Some near-field obstacles include obstacle remote sensing information representing obstacles around the autonomous mobile robot, that is, the grid type is obstacles identified by visual sensors. That is, only the obstacles in the map whose obstacle type is identified by images collected by visual sensors can be expanded, and areas of other obstacle types will not be expanded. The reason for this is that when marking obstacles, when other near-field obstacle information is received to mark the grid, since the size of the obstacle cannot be identified, in order to avoid missing obstacle information, enough grids are generally marked, so there is no need to expand them additionally. The map obtained in this way is more in line with the actual working scenario, and the machine can more easily find the direction to escape from the predicament.
[0111] Furthermore, when expanding, the grid type that meets the requirements can be deleted from the map, and the grid type can be lifted. According to multiple tests by staff, there may not be any insurmountable obstacles at the location where the lift is detected. Therefore, the grid type of lifted in the obstacle location map can be deleted, and the passable direction can be searched based on the map after deleting these obstacles, and an escape path can be generated. Compared with before deletion, this method increases the area on the map that is not covered by obstacles, and the resulting map is more consistent with the actual work scene, greatly increasing the probability of the lawn mower getting out of trouble.
[0112] In one embodiment of the present application, searching for a traversable direction based on the current position of the autonomous mobile robot and an obstacle location map may include: obtaining a search direction and search range for near-field obstacles from the obstacle location map; determining a direction within the search range that does not contain near-field obstacles as a traversable direction; the initial search direction is the direction opposite to the autonomous mobile robot's movement direction; and the search range is the area beyond the autonomous mobile robot's coverage on the map, as determined by the search direction. Specifically, when a lawn mower is detected to be trapped, a map of the area near the mower is first obtained. Then, with the mower's location as the center, obstacles within the area covered by the mower in the grid map are determined. For example, a circular area may be defined with a radius greater than the length of the mower's body and centered at the mower's center of gravity. Directions within the circular area within the grid map that do not contain near-field obstacles are determined as traversable directions.
[0113] Furthermore, when the lawn mower's body is divided into two parts with unequal lengths, the corresponding obstacle search range on the map can also be divided into two unequal parts, and obstacles in these two parts of the map can be searched separately. For example, the front half of the body covers 4 grids in the length direction of the map, the back half covers 8 grids in the length direction of the map, and the body covers 10 grids in the width direction. Therefore, in the map corresponding to the front half of the body, obstacle information covered by a semicircle with a radius of 8 grids (40 cm) centered on the lawn mower's location can be obtained; in the map corresponding to the back half of the body, obstacle information covered by a semicircle with a radius of 10 grids (40 cm) centered on the lawn mower's location can be obtained.
[0114] Furthermore, the search range can also be determined based on the search direction, which is determined to be beyond the area covered by the autonomous mobile robot on the map. The search direction used during the initial search is opposite to the robot's movement direction. Setting the search range can reduce the amount of data computation required when the lawn mower searches for an escape path. For example, in one embodiment, when the lawn mower is detected in distress, the linear velocity direction of the machine is first obtained. If the linear velocity direction is forward, the semicircular coverage area on the map, centered at the machine's center of gravity and with a radius of 10 grid cells in the rear half of the machine, is obtained as the search range.
[0115] In an embodiment of the present application, during the process of searching for a passable direction, when a passable direction exists in the initial search direction and search range, the initial search direction and search range are kept unchanged during the subsequent movement, and only the obstacle information in the initial search direction and search range is obtained. For example: when it is determined that there is a passable direction behind the lawn mower in the map when it is trapped, only the map behind it is obtained in the subsequent process of continuing to escape. By controlling the machine to move continuously in the same direction in this way, it is prevented that the lawn mower repeatedly switches the search direction and search range, resulting in the lawn mower sometimes moving forward due to the presence of a passable direction in front, and sometimes moving backward due to the presence of a passable direction behind the lawn mower, which causes the lawn mower to oscillate back and forth.
[0116] For obstacles in the map whose obstacle type is lifted, the search direction and search range can be determined based on the installation position of the lift sensor on the lawn mower, and the search direction and search range remain unchanged throughout the escape process. Specifically, when the installation position is behind the fuselage, the search direction is the direction of movement of the lawn mower, and the search range is the range covered by the map beyond the front half of the fuselage, that is, the lawn mower will move forward to escape; when the installation position is in front of the fuselage, the search direction is the direction opposite to the direction of movement of the lawn mower, and the search range is the range covered by the map beyond the front and rear half of the fuselage, that is, the lawn mower will move backward. Through this processing method, the lawn mower can pass the lift position as quickly as possible and escape from the predicament.
[0117] In one embodiment of the present application, when an autonomous lawn mower is detected as trapped, a map of the area surrounding the mower is obtained. If multiple unmarked areas exist, the mower is controlled to move to a closer area or direction based on the distance between the mower and the unmarked area to escape. Preferably, if there is a path forward, the mower moves directly forward.
[0118] In one embodiment, when an escape path cannot be generated, an area in the map is searched in the direction opposite to the initial search direction. When a passable direction is found, the lawn mower is controlled to move in that direction. As in the above embodiment, when the lawn mower is detected as trapped, the linear velocity direction of the machine body is first obtained. If the linear velocity direction is directly forward, the semicircular coverage area centered on the machine position in the map and with a radius of 10 grid cells in the rear half of the machine body is obtained as the search range. If a passable direction cannot be found within this search range, the search direction is switched to a semicircular coverage area with a radius of 8 grid cells in the front half of the machine body as the search range.
[0119] In one embodiment, when obstacles exist within the search range in the front and rear directions of the lawn mower on the map and no escape path can be generated, the search range is narrowed, but the search range is larger than the lawn mower body, and the passable direction is determined based on the narrowed search range.
[0120] In one embodiment, the traversable direction is adjusted accordingly as the mower's current position and the obstacle location map are updated. As the mower's current position is updated, the mower's movement direction is updated accordingly, as are the initial search direction and search range. The obstacle location map is also continuously updated with the collected near-field obstacle information. However, since the selected traversable direction is the direction closest to the mower, the traversable direction remains essentially unchanged unless the mower's position or the obstacle location map changes significantly.
[0121] In one embodiment, while controlling an autonomous mobile robot to move in a traversable direction, if the time it takes to detect an obstacle is less than a preset time threshold, the robot is controlled to continue searching for a traversable direction in a direction consistent with the initial search direction. If the time exceeds the preset time threshold, the robot is controlled to re-search for a traversable direction according to the method described in this application. In this case, the traversable direction may be opposite to the initial search direction. Because obstacles may be temporary or misdetected by the robot, the robot is controlled to re-search for a traversable direction only after a long period of time or multiple detections of new obstacles, thereby ensuring the stability of the robot's movement process.
[0122] In another embodiment, searching for a traversable direction based on the current position of the autonomous mobile robot and the obstacle location map may include: drawing multiple sectors D on the map centered on the center of gravity of the autonomous mower, assigning a value to each sector based on whether the sector passes through an obstacle, and assigning no value if the sector does not pass through an obstacle. Controlling the mower to move in the direction of the unassigned sectors. Specifically, taking the dilemma scenario shown in FIG4 as an example, multiple rays as shown in FIG5 are drawn at equal angles within a 360-degree range around the mower on the map, with two adjacent rays defining a sector, resulting in N sectors. Sector values are determined for each direction based on whether an obstacle exists in the grid covered by each sector. If an obstacle exists, the sector value for that direction is marked as 1. Furthermore, the distance of the nearest obstacle to the mower in the map grid covered by each sector in each direction is obtained, and the number of sectors to be expanded is determined based on the distance 2D obtained for each sector, the vehicle body width d, and the sector resolution (2pi / N). At this time, the number of sectors that need to be expanded for each sector is calculated as: (Arcsin(d / D)) / (2pi / N). The number of sectors that need to be expanded represents the number of sectors that will be covered when the lawn mower moves to that sector. For example, if the calculated number of sectors is 2, the values of the four sectors near the current sector can be marked as 1. The direction of the sector ray with a sector value of 0 is obtained. This direction or the sector area can be a passable direction or area for the lawn mower to escape. Furthermore, the direction of the sector ray can be the direction closest to the current moving direction of the lawn mower. Compared with the method of simulating the grids covered when the lawn mower reaches the obstacle by simulating the area covered by the lawn mower body when it is located at each obstacle in the above embodiment, the sector simulation method does not need to simulate all the grids covered by the machine when it moves to the obstacle. Therefore, while ensuring that the obstacles around the lawn mower are simulated and covered, there is no redundancy and the algorithm is simpler. Therefore, when the lawn mower detects that it is trapped, it can escape the direction of the predicament more quickly.
[0123] In another embodiment of the present application, when an obstacle exists, the sector value in that direction is marked as 1 and the fault type is recorded. When the sector is expanded, only the obstacle type identified by the image collected by the visual sensor can be expanded, and the areas of other types are not expanded. The reason for this is that other fault types have already been marked with enough grids, so there is no need to expand them additionally. The obstacle area marked in this way is more consistent with the actual situation of the obstacle, and the machine can more easily find the direction to escape from the predicament.
[0124] In one embodiment of the present application, when the autonomous mobile robot is unable to search for a passable direction, the autonomous mobile robot is controlled to ignore the obstacle location map and adjust its posture. The posture adjustment method may include but is not limited to at least one of the following: forward, backward, and rotation; based on the current position of the autonomous mobile robot after the posture adjustment and the obstacle location map, a passable direction is searched. When the lawn mower detects that it is in trouble, if the method in the above embodiment is used in combination with the map and it is unable to search for a passable direction forward or backward, the lawn mower can be controlled to rotate directly on the spot and search for a passable direction at the rotated position; when the lawn mower detects that it is in trouble, if the method in the above embodiment is used to move forward in an arc for a long time and it can still receive near-field obstacle information such as lifting, collision, or torque, the lawn mower can be controlled to move forward a certain distance and rotate, and search for a passable direction at the rotated position.
[0125] Furthermore, after the lawn mower adjusts its posture, the near-field obstacle recognized in real time by the visual sensor when the lawn mower is at its current position is re-acquired. When the recognized near-field obstacle is inconsistent with the corresponding visually recognized obstacle position in the obstacle position map, the map is updated according to the near-field obstacle recognized in real time by the visual sensor, and a passable direction is searched according to the current position of the lawn mower after the posture adjustment and the updated obstacle position map.
[0126] After determining a passable direction, the angular velocity and linear velocity of the autonomous mobile robot are determined based on the angular difference between the autonomous mobile robot's moving direction and the passable direction, and the autonomous mobile robot is controlled to move in the passable direction according to the angular velocity and linear velocity, wherein the angular velocity is positively correlated with the angular difference, and the linear velocity is negatively correlated with the angular difference. For example, for the above-mentioned sector embodiment, if the distance between the current direction of the lawn mower and the target ray direction is greater than N / 4, the linear velocity is set to 0.03m / s and the angular velocity is set to 0.025m / s×d; if the distance between the current direction of the lawn mower and the target ray direction is less than N / 4 and greater than N / 6, the linear velocity is set to 0.05m / s and the angular velocity is set to 0.02m / s×d. This speed setting method allows the lawn mower to turn to the target position or target direction more quickly when the position change is relatively small.
[0127] Furthermore, when the passable direction is rearward, the lawn mower is controlled to escape by moving in an arc or backward according to the moving angular velocity and the moving linear velocity; when the passable direction is forward, the lawn mower is controlled to escape by moving in an arc or straight according to the moving angular velocity and the moving linear velocity.
[0128] In one embodiment of the present application, when controlling an autonomous mobile robot to move in a passable direction, it is determined whether it can rotate based on a map of obstacle positions around the autonomous mobile robot; if it can rotate, it is controlled to rotate, and whether the current direction is a passable direction is confirmed based on the position after rotation.
[0129] The robot determines whether it can rotate by calculating whether there are obstacles in the map covered by the robot after rotating according to a preset angle. For example, a lawn mower simulates moving to multiple preset angles to determine whether there are obstacles in the area covered by the map. If there are no obstacles, the robot can rotate. The preset angles can be 30, 60, 90, 120, 150, or 180 degrees counterclockwise, or 30, 60, 90, 120, 150, or 180 degrees clockwise. The robot is preferentially searched in a counterclockwise direction, which aligns with the mower's edge direction. When a rotational direction is found, the robot is controlled to rotate in that direction. Based on the rotated position, the robot searches for a traversable direction. Once a traversable direction is found, the robot is controlled to travel in a straight line in that direction. Whether the current direction is traversable can be confirmed in real time during rotation, or the robot can be stopped after each rotation of a certain angle to confirm whether the current direction is traversable. Preferably, the initial rotation direction is determined based on the mower's edge-along direction. If the edge-along direction is counterclockwise, the initial direction is preferably counterclockwise. When the mower rotates in this direction, it can quickly edge-along. If counterclockwise rotation is not possible, the mower switches to determining whether clockwise rotation is possible. If neither direction is possible, the mower continues to move in the previously searched passable direction.
[0130] In one embodiment of the present application, while controlling the rotation of an autonomous mobile robot, obstacle remote sensing information collected in real time by an obstacle remote sensing sensor is used to determine whether there is a near-field obstacle ahead. The autonomous mobile robot is equipped with an obstacle remote sensing sensor configured to identify obstacle remote sensing information surrounding the autonomous mobile robot. When it is determined that no near-field obstacle exists, the direction in question is determined to be a traversable direction, and the autonomous mobile robot is controlled to move in that direction. For example, while a lawn mower is rotating, images collected by a visual sensor are used to determine whether there is an unobstructed area ahead that can accommodate the width of the mower. If so, the mower is controlled to move in that direction. Specifically, visually determining whether there is an unobstructed area ahead that can accommodate the width of the mower includes segmenting an image to extract a roughly rectangular area from the image (with the desired straight-line range on one side and the width of the mower or slightly less than the width of the mower on the other side). Based on the pixels in the extracted image, a preset ratio of the obstacle area to the entire area is calculated. If the ratio is below a preset ratio (e.g., 5%), it indicates that there are no obstacles and the autonomous mobile robot can pass through.
[0131] Furthermore, the presence of near-field obstacles ahead can be determined using real-time obstacle remote sensing information collected by the obstacle remote sensing sensor and the obstacle location map. The search range is the sector where the mower's head is located during rotation, and the obstacle location map is used to determine whether this sector is a passable direction. Furthermore, the imagery collected by the visual sensor is used to identify whether there is an unobstructed area ahead that can accommodate the width of the mower. If the two conclusions are consistent, the mower is controlled to move in that direction.
[0132] The multiple embodiments of obtaining a passable direction described in this application, such as: deleting some grid types, selecting search directions and search ranges, maintaining search directions, and switching search directions when an escape path cannot be generated, switching search directions when obstacles are detected while driving in a passable direction, and sectors, etc., these embodiments can be applied separately to the obstacle location map to search for a passable direction, or they can be used in combination. For example, the deletion of raised grid types, the selection of search directions and search ranges, the maintenance of the initial search direction, etc. can be combined with a sector solution and applied to the obstacle location to determine the passable direction. This application does not limit this.
[0133] In another embodiment of the present application, based on the current position of the autonomous mobile robot and the obstacle position map, searching for a passable direction may also include: randomly generating multiple moving directions with the position of the lawn mower as the center on the map, simulating the positions it will reach after moving along these directions, and identifying whether these positions overlap with obstacle points (grid value is 1); selecting non-overlapping moving directions; calculating evaluation parameters for each non-overlapping moving direction, the evaluation parameter being related to the distance between the position reached by the lawn mower after the simulated movement and the obstacle point, for example: 80 / distance squared, and selecting the direction with the smaller evaluation parameter as the moving direction of the machine to escape.
[0134] In one embodiment of the present application, if the lawn mower is detected to be in trouble, it can be controlled to continue working. If it is detected that it still has an abnormality such as slipping, collision, lifting, or motor stalling during operation, the controller will stop working to avoid over-cutting certain areas.
[0135] After the autonomous mobile robot is detected to be trapped, the autonomous mobile robot is determined to have successfully escaped when it travels at least a first preset distance along the escape path. For example, after the autonomous mobile robot is detected to be trapped, the autonomous mobile robot is determined to have successfully escaped when it travels at least 15 cm after identifying that a real collision has been resolved.
[0136] After the mower is determined to have escaped successfully, the area on the map that was covered by the mower when it was detected as an obstacle is marked. During subsequent path planning, the route is replanned based on the re-marked map to follow the work area boundary or the planned path. This prevents the mower from accidentally moving into the same area during subsequent movements, which could reduce work efficiency.
[0137] After the autonomous mobile robot is determined to have successfully escaped, if it is detected that the autonomous mobile robot has moved to a second preset distance from the trapped location, obstacles marked in the map by sensors installed on the autonomous mobile robot due to a fault detected are deleted. For example, after the lawn mower is determined to have successfully escaped, if it is detected that it has moved to 5 meters from the trapped location, obstacles marked due to abnormal conditions such as camera collision, torque change, or universal wheel lifting are deleted. When the map accuracy is low, the accuracy of the obstacle mark position will also decrease as the lawn mower moves longer, so it can be deleted. Because the visual sensor updates data as it captures images, the obstacles it has marked can be retained.
[0138] In one embodiment of the present application, in the internal right-angle scenario shown in Figure 1, when multiple collisions are detected, if the above-mentioned sector and search range embodiment is adopted, it can be found that 2 / 3 of the sectors in the second half are marked as 1, and 1 / 2 of the other half are marked as 1. When this scenario is identified, the machine can be controlled to travel at a faster angular velocity and a slower linear velocity, so as to leave the built-in corner faster.
[0139] In an embodiment of the present application, near-field obstacle information transmitted by at least one sensor is received, and based on the near-field obstacle information, a map of obstacle locations surrounding the autonomous mobile robot is generated. In response to receiving near-field obstacle information multiple times within a limited time or space, an escape path is generated. Generating the escape path includes searching for a navigable direction based on the autonomous mobile robot's current position and the obstacle location map. Compared to detecting a robot in distress and then controlling it to randomly swerve in various directions to escape, the method described in this application allows the lawn mower to escape quickly, improving work efficiency.
[0140] The following describes an embodiment of the present application using an application scenario as an example. In the scenario shown in FIG4 , a lawn mower collides multiple times with multiple obstacles 10 within a confined space during travel, indicating that the lawn mower is in a predicament. Based on the received near-field obstacle information, a map E of obstacle locations surrounding the lawn mower is generated. This map is referred to as a local map E. Knowing that the lawn mower's most recent collision occurred in direction A during diagonal movement, the search range is used with the rear half of the machine as the semicircle, resulting in FIG6 . The sectors shown in FIG6 are marked based on the local map E, with the positions marked as 1 in each sector marked as 1 in the local map E. The obstacles in the sectors are expanded according to (Arcsin(d / D)) / (2pi / N), resulting in a sector map with values as shown in FIG6 . The direction B corresponding to the sector closest to direction A in the semicircle with a value of 0 is obtained as the traversable direction. The linear velocity and angular velocity of the machine's backward arc movement are determined based on the angular difference between direction A and the traversable direction B. The lawn mower is controlled to traverse in the traversable direction B as shown in FIG6 . During the process of traveling in the direction B, the obstacle position map determines that the lawn mower can rotate, so the lawn mower is controlled to rotate. Based on multiple positions during the rotation, it is determined that the lawn mower is passable in the forward direction corresponding to direction C. The lawn mower is controlled to move in direction C as shown in Figure 6. After moving in a straight line in direction C for a period of time, it is determined that the lawn mower has successfully escaped.
[0141] In addition, an embodiment is provided, which is an autonomous mobile robot, comprising:
[0142] processor;
[0143] a memory for storing processor-executable instructions;
[0144] A processor is used to execute the above method.
[0145] The above methods can all be executed in the controller of the lawn mower. When obstacle detection is involved, the relevant information can be detected by a sensor, and after the information is detected, it is sent to the controller to control it to execute the corresponding method. This application will not be elaborated here.
[0146] In addition, an embodiment is provided, which is a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is used to execute the above method.
[0147] Another embodiment of the present application further provides a control method for an autonomous mobile robot as shown in FIG8 . In this embodiment, the to-be-traveled area and the traversable direction have substantially the same meaning, and includes the following steps:
[0148] S1, detecting whether the autonomous mobile robot meets the conditions for escaping;
[0149] S2, if the conditions for escaping are met, detecting whether there are obstacles within a range corresponding to the area to be driven in the image in front of the autonomous mobile robot;
[0150] S3: If there is no obstacle in the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to drive out of the area to be driven to escape.
[0151] The autonomous mobile robot includes a visual sensor for capturing images in front of the autonomous mobile robot. The visual sensor can be, for example, a camera. Through the above method, when the autonomous mobile robot meets the conditions for escape, the image captured by the visual sensor is used to detect whether there are obstacles in the area to be driven in front of the autonomous mobile robot. The autonomous mobile robot is then controlled to escape from the obstacle-free area. Visual recognition enables the autonomous mobile robot to more specifically escape from obstacles, effectively improving its efficiency in escaping complex areas.
[0152] The conditions for escaping from distress include at least one of the following:
[0153] The autonomous mobile robot has a collision greater than the preset number;
[0154] The driving wheels of the autonomous mobile robot are blocked;
[0155] The lifting sensor detects that the autonomous mobile robot is lifted;
[0156] When the autonomous mobile robot is mowing in situ, the actual rotation speed of the cutter disc is less than or equal to the preset rotation speed;
[0157] The image acquired by the vision sensor remains unchanged for a preset time period.
[0158] Optionally, the autonomous mobile robot meets different conditions for escape, indicating that the device is in different trapped scenarios. Among these, the autonomous mobile robot experiencing more than a preset number of collisions, the autonomous mobile robot's drive wheels becoming stuck, or the lift sensor detecting that the autonomous mobile robot is being lifted typically occur in narrow areas or those with multiple pillars, i.e., multiple collision trapped scenarios. For a lawn mower, when the actual cutterhead speed is less than 55% while the autonomous mobile robot is mowing in place, this typically occurs in dense grass, i.e., a dense grass trapped scenario. Specifically, the triggering method includes: controlling the machine to mow in place when the actual cutterhead speed is less than or equal to a preset speed and the actual cutterhead torque is greater than 90%. If the actual cutterhead speed is less than 55% while mowing in place, it indicates that the machine has encountered dense grass. The image captured by the visual sensor remaining unchanged for a preset period typically occurs in a slipping scenario, where the drive wheels are spinning despite no apparent movement of the machine, i.e., a slipping trapped scenario. Optionally, the autonomous mobile robot's trapped scenario can be determined based on the conditions it meets for escape, and different escape strategies can be implemented based on different trapped scenarios. The device may or may not work when trapped.
[0159] Among them, the area to be driven is an area within a certain range in front of the autonomous mobile robot, which is an area in the real environment space. When detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot, the area to be detected in the image is the area projected into the image by the area to be driven.
[0160] In order to improve the success rate of escape, the area to be driven can be determined within two distance ranges. Please refer to Figure 9. One layer is to determine the area to be driven within the range of R1, and the other layer is to determine the area to be driven within the range of R2. Since the ideal obstacle-free area cannot be detected within the range of R1, it may be because there is an obstacle at the boundary that is almost the same size as the width of the equipment body, and the obstacle may not affect the movement of the equipment. Therefore, the range can be appropriately narrowed and the area to be driven can be determined within the range of R2, which can more effectively improve the success rate of the equipment's escape.
[0161] The area to be driven can be a nearly rectangular area within a certain range in front of the autonomous mobile robot. In order to enable the autonomous mobile robot to drive and escape relatively smoothly, one side of the area to be driven can be selected to be larger than the body width of the autonomous mobile robot, and the other side is a second preset distance. The second preset distance is the distance in the driving direction of the autonomous mobile robot, corresponding to R1. In order to enable the autonomous mobile robot to drive to a certain extent in complex areas to ensure escape, one side of the area to be driven can be selected to be smaller than the body width or equal to the body width, and the other side is a third preset distance. The third preset distance is the distance in the driving direction of the autonomous mobile robot, corresponding to R2, wherein the third preset distance is smaller than the second preset distance. In actual implementation, the above two different ranges of areas to be driven can be used at different stages of escape to effectively improve the success rate of the autonomous mobile robot in escaping.
[0162] Optionally, the method of the present application further includes:
[0163] If there is an obstacle in the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to rotate;
[0164] Enter the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot.
[0165] Optionally, before rotating, the autonomous mobile robot may first perform a backward operation to move away from the detected obstacle. If an abnormal situation occurs during the backward operation, such as detecting an obstacle, the rotation operation may be initiated.
[0166] Optionally, if there is an obstacle within the range corresponding to the area to be driven in the image, the autonomous mobile robot can be controlled to rotate in any direction; the machine's rotation direction can also be controlled based on whether the collision detected when trapped is a virtual collision or a physical collision; the machine's rotation direction can also be determined based on the position of the obstacle in the image captured by the camera. For example, if a virtual collision has already occurred to escape the trap, if a virtual collision occurs again, the machine can be controlled to continue rotating in the same direction as the last time it escaped the trap. Of course, considering issues such as stability during the machine's operation, before rotating, the machine can be controlled to remain stationary for 500ms, and subsequent operations can be performed after the machine's performance stabilizes.
[0167] Optionally, if there is an obstacle within a range corresponding to the area to be driven in the image, controlling the autonomous mobile robot to rotate includes:
[0168] If there is an obstacle within the range corresponding to the area to be driven in the image, determining whether the autonomous mobile robot has rotated a first preset angle in the current rotation direction;
[0169] If the first preset angle has not been rotated, the autonomous mobile robot is controlled to rotate according to the current rotation direction.
[0170] Optionally, if an obstacle exists within a range corresponding to the area to be driven in the image, controlling the autonomous mobile robot to rotate further includes:
[0171] If the first preset angle has been rotated, the area to be driven is reduced, and the autonomous mobile robot is controlled to rotate in the opposite direction of the current rotation direction.
[0172] The area to be driven before shrinking has one side greater than the width of the autonomous mobile robot, and the other side is a second preset distance. The area to be driven after shrinking has one side less than the robot by a preset width or equal to the robot's width, and the other side is a third preset distance, which is less than the second preset distance.
[0173] Among them, the autonomous mobile robot first rotates at a low speed in a certain direction to find a passable exit (for example, forward rotation, the rotation speed can be 10cm / s), and uses the visual sensor to detect whether there is an obstacle-free area in the front range R1 (for example, 30cm) that can accommodate the machine's movement. The detection width at this time can be slightly larger than the width of the machine to ensure that the machine can escape smoothly. That is, one side of the area to be traveled at this time is larger than the body width of the autonomous mobile robot, and the other side is a second preset distance. If there is an obstacle-free area in the R1 range, the autonomous mobile robot is controlled to walk in the direction of the obstacle-free area, and after walking continuously for a certain period of time (for example, 4s), it is confirmed that it has exited the trapped area, successfully escaped, released the escape mode, and re-entered the working mode. If an obstacle appears again in the R1 range, repeat the above strategy to escape.
[0174] If the autonomous mobile robot still fails to detect an obstacle-free area within the forward range R1 after rotating through a first preset angle (e.g., 270 or 360 degrees, where the visual sensor's field of view has already completed a full rotation), this indicates that there is no obstacle-free area within R1. In this case, the detection radius is reduced to R2 (e.g., 10 cm), and the autonomous mobile robot is controlled to rotate in the same or opposite direction to detect an obstacle-free area. The detection width can be the width of the autonomous mobile robot or slightly less (e.g., the width of the robot minus the width of one wheel) to ensure the robot can escape. That is, one side of the area to be driven is greater than the width of the autonomous mobile robot, and the other side is a third preset distance. In practical implementation, when reducing the detection radius to R2, it is preferable to switch to the reverse direction of rotation because the 30 cm range has already been detected before the reversal, which can improve detection efficiency. If the robot detects an obstacle, non-grass, or magnetic strip during the first preset angle rotation, preventing it from continuing to rotate, the robot can initiate reverse rotation or reduce the detection radius to R2 and initiate reverse rotation.
[0175] If there is an obstacle-free area within R2, the autonomous mobile robot is controlled to move in the direction of the obstacle-free area. After walking for a certain period of time (for example, 4 seconds) or a certain distance, it is confirmed that it has exited the trapped area and successfully escaped. The escape mode is released and the robot re-enters the working mode. If an obstacle appears again within R2, the above strategy can be repeated within R1 or R2 to escape.
[0176] When an obstacle is detected during the detection of the obstacle-free area, the obstacle position and its angle can be recorded.
[0177] The third preset distance can be selected based on the performance of the autonomous mobile robot. For example, if the performance of the autonomous mobile robot determines that when the autonomous mobile robot detects an obstacle or boundary within a 10 cm range, the device can stop before encountering the obstacle or boundary based on the image recognition and processing response time and the machine's reaction time. In this case, the third preset distance can be selected as 10 cm or slightly greater than 10 cm. This ensures the safety of the device during movement.
[0178] The process for detecting whether there are obstacles within the area of the autonomous mobile robot's travel in an image ahead of it is to segment the image and extract a roughly rectangular area (30 cm / 10 cm on one side, with the width of the robot or slightly less than the width of the robot on the other side) from the image. Based on the pixels in the captured image, the area of the obstacle is calculated to determine whether it occupies less than a preset percentage (e.g., 5%) of the total area. If so, the area indicates that there are no obstacles and the autonomous mobile robot can pass through.
[0179] The first problem an autonomous mobile robot faces when it detects it's trapped is that continuing in its original direction will lead to impediment. Even after achieving movement through certain strategies, it may find itself stuck again. Therefore, when an autonomous mobile robot detects it's trapped, it first addresses this impediment. Solutions include retreating, rotating, and then advancing. The robot can retreat before rotating, and when retreating, it should retreat less than one unit length at a time to ensure safety. After retreating, if the device isn't in an abnormal state like stalled or lifted, the robot can be controlled to rotate. Otherwise, the robot can be controlled to advance and then rotate. Of course, in practice, this doesn't necessarily require retreating before rotating; it can also rotate directly in place. During the rotation process, the aforementioned strategy is used to locate an obstacle-free area.
[0180] In an embodiment of the present application, when the machine retreats, it can adopt a straight-line retreat or an arc-shaped retreat method, and can also choose to adopt an arc-shaped retreat or a straight-line retreat retreat strategy according to the collision type (virtual collision and physical collision). When an abnormality is detected during the machine retreat, the machine can be controlled to rotate in an attempt to escape from the predicament. When a physical collision occurs between the machine and an obstacle in direct contact and it is determined that the machine is trapped, since this direct contact collision may cause the machine position to shift, the position of the machine after the shift may be a position where it can escape. Therefore, when the machine detects that it is trapped, it can adopt a straight-line retreat method. When a physical collision does not occur between the machine and the obstacle in direct contact and it is determined that the machine is trapped, for example, when the camera on the machine captures an obstacle in the walking direction and needs to escape, it can adopt an arc-shaped retreat method. Since the camera captures a wide range, the arc-shaped retreat method increases the randomness of the retreat and improves the probability of successful escape.
[0181] During the image acquisition process, based on the rotation speed of the autonomous mobile robot and the image acquisition speed of the visual sensor, an image can be acquired once when the device rotates a certain angle, so that images can be acquired at equal intervals (for example, an image is acquired approximately every 4 degrees of rotation). When the autonomous mobile robot rotates faster or the acquisition speed is faster, the angular interval of the acquired images will be shortened.
[0182] Optionally, if there is no obstacle within the range corresponding to the area to be driven in the image, after controlling the autonomous mobile robot to drive out of the area to be driven to escape, the method further includes:
[0183] Determining whether the distance traveled by the autonomous mobile robot from the area to be traveled is less than a first preset distance;
[0184] If the distance is less than the first preset distance, the autonomous mobile robot is determined to meet the conditions for escaping, and the number of escape attempts is accumulated;
[0185] The method of the present application also includes:
[0186] Before controlling the autonomous mobile robot to rotate, determine whether the number of escape attempts corresponding to the current rotation direction has reached a preset number;
[0187] If the preset number of times is reached, the autonomous mobile robot is controlled to rotate and pause to acquire images;
[0188] When the autonomous mobile robot rotates to the second preset angle, it restarts acquiring images and resets the number of escape attempts to zero.
[0189] Among them, the first preset distance can be equal to or less than the third preset distance. After determining that there are no obstacles in the current area to be driven, the driving distance is less than the area to be driven, indicating that there is a possibility of misidentification. For example, when the fuselage is tilted, it may cause the passable area to be misidentified. In order to avoid the defect of low efficiency caused by multiple attempts to escape from the predicament and failure, the autonomous mobile robot is forced to rotate to a second preset angle (for example, a random angle of 20-70 degrees or 25-75 degrees) every time the number of attempts to escape reaches a preset number (for example, 4 times), and then re-collect images for detection. When the machine attempts to escape more than a certain number of times, for example 8 times, it means that the machine has failed to escape, and the machine is controlled to stop.
[0190] Optionally, the method of the present application further includes:
[0191] If there is an obstacle in the range corresponding to the area to be driven in the image, the area to be driven is reduced;
[0192] Enter the step of detecting whether there are obstacles within the range corresponding to the area to be driven in the image.
[0193] Among them, when there is an obstacle within the range corresponding to the area to be driven in the image, the area to be driven can be narrowed without rotating and / or changing the rotation direction, and further detection can be made as to whether there is an obstacle within the range corresponding to the area to be driven in the image in an attempt to escape.
[0194] The above escape strategy is applicable to multiple collision trapped scenarios. In multiple collision trapped scenarios, using smaller angular intervals to capture images and / or a smaller rotation speed can prevent the machine from missing the exit that can help it escape from the predicament.
[0195] Optionally, the method of the present application further includes at least one of the following:
[0196] If the autonomous mobile robot is currently trapped in dense grass, before the autonomous mobile robot is controlled to rotate for the first time, the cutter head of the autonomous mobile robot is controlled to stop and move backward. When the autonomous mobile robot moves backward a preset distance, the cutter head of the autonomous mobile robot is controlled to start rotating.
[0197] If the autonomous mobile robot is currently in a slipping and trapped scenario, the autonomous mobile robot is controlled to rotate at a first speed for the first time, and a different speed is used in each rotation, and the first speed is less than a preset speed.
[0198] In the case of an AMR trapped in dense grass, in addition to the aforementioned strategies for multiple collision scenarios, before the AMR is initially controlled to rotate, the AMR will be controlled to stop the cutterhead and move backward. After moving backward a preset distance, the AMR will be controlled to start rotating. Unlike multiple collision scenarios, dense grass typically covers large areas. When dense grass is detected, the AMR will stop the cutterhead and move backward. After moving to a certain position, the AMR will start rotating at a large angle, which can improve work efficiency.
[0199] When the autonomous mobile robot is currently in a slipping and trapped scenario, in addition to the above-mentioned strategy applicable to the multiple collision trapped scenario, the autonomous mobile robot is controlled to rotate for the first time at a first speed, and a different speed is used for each rotation, and the first speed is less than the preset speed. Thus, by starting at a low speed and frequently switching the rotation speed of the drive wheel, the machine generates relative motion with the ground due to inertia, thereby increasing the possibility of getting out of trouble, and preventing the defects of the machine being unable to get out of trouble when using the same speed or using a higher speed causing the grass to bend and the pit to expand, making it even more difficult to get out of trouble. In the slipping and trapped scenario, the autonomous mobile robot can also use the actions of retreating, rotating, and moving forward to get out of trouble. When retreating, the moving speed is a fixed value. Each time it rotates or each time it moves forward, the speed of rotation or forward can be adjusted (for example, 10cm / s->20cm / s->30rpm / s). When initially getting out of trouble, both rotation and forward speeds are used at a lower speed, which is slowly increased in the later stage, and then reduced again to ensure that the machine can get out of trouble. Specifically, when the device slips and escapes, it moves backward and then rotates at a speed of 10 cm / s. If it can rotate, it detects the obstacle-free area during the rotation process and controls the device to move in the direction of the obstacle-free area after rotation. If it can move for a preset time (for example, 4 seconds), it indicates that the escape is successful. If it cannot rotate, it attempts to move forward. If it can move forward, it indicates that the escape is successful after the device is controlled to move for another preset time (for example, 8 seconds). If it cannot move forward, it continues to move backward, changes the speed of forward movement or rotation, and repeats the above process.
[0200] Optionally, if there is no obstacle within a range corresponding to the area to be driven in the image, controlling the autonomous mobile robot to drive out of the area to be driven to escape, including:
[0201] If there are no obstacles within the range corresponding to the area to be driven in the image and the vehicle is currently in a preset trapped scene, determining whether the rotation angle in the current rotation direction is greater than a third preset angle;
[0202] If the rotation angle is greater than the third preset angle, the autonomous mobile robot is controlled to drive out of the waiting area to escape.
[0203] Optionally, the method further comprises:
[0204] If the rotated angle is less than or equal to the third preset angle, controlling the autonomous mobile robot to rotate in the current rotation direction;
[0205] Enter the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot.
[0206] Among them, the preset scenes include a dense grass trapped scene and a slipping trapped scene, and the third preset angles corresponding to the dense grass trapped scene and the slipping trapped scene may be different or the same. In the dense grass trapped scene, since dense grass is generally a large area, when the control device is rotated to a random angle greater than 60° (the third preset angle), if there is an obstacle-free area and then move forward to exit the trapped area, the escape efficiency can be improved. In the slipping trapped scene, the device may be in a pit. When the control device is rotated to a random angle of 90-180 degrees (the third preset angle), if there is an obstacle-free area and then move forward to exit the trapped area, the device can be prevented from being trapped in the pit. Due to the small rotation angle, the device is restored to its original angle again due to the slope and its own weight during the forward process, resulting in a defect of low escape efficiency.
[0207] Optionally, after controlling the autonomous mobile robot to drive out of the waiting area to escape, the method further includes:
[0208] Determine whether the escape was successful;
[0209] If the vehicle is successfully rescued, the autonomous mobile robot is controlled to start working.
[0210] Based on the above content, the following describes the escape strategies that can be used in scenarios of multiple collisions, dense grass, and slipping.
[0211] (1) Multiple collision and entrapment scenarios
[0212] Eligibility criteria include at least one of the following:
[0213] The autonomous mobile robot has a collision greater than the preset number;
[0214] The driving wheels of the autonomous mobile robot are blocked;
[0215] The lift sensor detects that the autonomous mobile robot is lifted.
[0216] Multiple collision trapping scenarios often occur in confined areas. These scenarios can be categorized as contact collisions (physical collisions) or non-contact collisions (virtual collisions) based on whether the autonomous mobile robot directly collides with an obstacle. A contact collision is when the robot escapes due to a direct collision with an obstacle. A non-contact collision is when the robot escapes without a direct collision with an obstacle. For example, an escape is triggered by the early detection of an obstacle by sensors such as cameras.
[0217] In an embodiment of the present application, when the machine has a contact collision such as a collision, lifting, or drive wheel jam that meets the requirements, it can be determined that it is trapped and an escape action is triggered. Whether the machine is trapped and an escape action is triggered can be determined based on whether a collision, lifting, or jam that meets the duration requirements is detected, or whether a collision, lifting, or jam that meets the number requirements and time interval requirements is detected. Specifically, when the machine has three consecutive collisions and the interval between each collision is less than 10s, or when the autonomous mobile robot detects a collision for 3 seconds, it can be considered that the machine is trapped and it is necessary to trigger an escape strategy corresponding to the multiple collision and trapped scenario. When 20 collisions are detected within 5 minutes, it is determined that the scene encountered by the machine at this time is complex and may not be able to be handled by itself, so it will be controlled to shut down and an alarm will be issued to remind the user to assist. Furthermore, in order to avoid the machine from falsely detecting a collision, it will only be determined that a collision has occurred if a collision is detected within 0.1s. When the machine is lifted three times in a row and the interval between each lift detection is less than 10s, or when the autonomous mobile robot detects lifts for 3 consecutive seconds, the machine can be considered trapped and the escape strategy corresponding to the multiple collision trapped scenario needs to be triggered. Furthermore, to avoid the machine from falsely detecting lifts, it is only when a collision is detected within 0.2s that it is determined that a collision has occurred. When the machine is stalled three times in a row and the interval between each stall detection is less than 10s, or when the autonomous mobile robot detects stalls for 3s, the machine can be considered trapped and the escape strategy corresponding to the multiple collision trapped scenario needs to be triggered (stalls include left wheel stalls or right wheel stalls). Furthermore, to avoid the machine from falsely detecting stalls, it is only when a stall is detected within 3s that it is determined that a collision has occurred. When the machine detects three consecutive collisions, lifts or stalls, it can also be considered trapped and the escape strategy corresponding to the multiple collision trapped scenario needs to be triggered. Of course, the duration and number of times used in the above embodiment to determine that the machine is trapped are only examples, and can be adaptively adjusted as the machine's detection accuracy is improved to improve the machine's response time to initiate the escape strategy. For example, the autonomous mobile robot can adjust the detection of collisions within 3 seconds to the detection of collisions within 2 seconds. This application is not limited to this. When any of the above-mentioned contact collisions, such as collision, lifting, and drive wheel stalling, is detected three times in a row, the machine can be determined to be trapped, thereby triggering the escape strategy corresponding to the multiple collisions.
[0218] In an embodiment of the present application, when a non-contact collision that meets the requirements occurs in a machine, it can be determined that it is trapped and an escape action can be triggered. When the autonomous mobile robot is detected to have made two turns in a small area within 40 seconds, it can be determined that it is trapped (for example, trapped in a small area) and an escape strategy needs to be initiated. Specifically, whether the machine has made two turns can be determined by detecting whether the yaw angle (Yaw) of the machine has changed by 540 degrees within 40 seconds. Of course, it can also be determined that it is trapped when a certain angle greater than one turn is detected, and this application is not limited to this. When the machine is detected to have made multiple rotations in place and multiple collisions within a short distance, it can be determined that the machine is trapped. For example, the machine is preset with benchmark values for rotations in place and collisions within a short distance. The machine's actions can be scored based on the frequency of the machine's actual rotations in place and the frequency of collisions actually detected. When the score is greater than a preset threshold, the machine is determined to be trapped (for example, trapped in a narrow passage), and an escape strategy corresponding to the multiple collisions can be triggered.
[0219] When a non-contact collision or a contact collision as described above is detected, it can be determined that the machine is trapped, thereby triggering an escape strategy corresponding to multiple collisions.
[0220] Detect whether the autonomous mobile robot meets the conditions for escaping; if the conditions for escaping are met, detect whether there are obstacles in the range corresponding to the area to be traveled in the image in front of the autonomous mobile robot; if there are no obstacles in the range corresponding to the area to be traveled in the image, control the autonomous mobile robot to drive out of the area to be traveled to escape.
[0221] If there is an obstacle in the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to rotate; and the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot is entered.
[0222] If there is an obstacle within the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to rotate, including:
[0223] If there is an obstacle within the range corresponding to the area to be driven in the image, determining whether the autonomous mobile robot has rotated a first preset angle in the current rotation direction;
[0224] If the first preset angle has not been rotated, the autonomous mobile robot is controlled to rotate according to the current rotation direction.
[0225] If there is an obstacle within the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to rotate, and further includes:
[0226] If the first preset angle has been rotated, the area to be driven is reduced, and the autonomous mobile robot is controlled to rotate in the opposite direction of the current rotation direction.
[0227] Optionally, if there is no obstacle within the range corresponding to the area to be driven in the image, after controlling the autonomous mobile robot to drive out of the area to be driven to escape, the method further includes:
[0228] Determining whether the distance traveled by the autonomous mobile robot from the area to be traveled is less than a first preset distance;
[0229] If the distance is less than the first preset distance, the autonomous mobile robot is determined to meet the conditions for escaping, and the number of escape attempts is accumulated;
[0230] The method also includes:
[0231] Before controlling the autonomous mobile robot to rotate, determine whether the number of escape attempts corresponding to the current rotation direction has reached a preset number;
[0232] If the preset number of times is reached, the autonomous mobile robot is controlled to rotate and pause to acquire images;
[0233] When the autonomous mobile robot rotates to the second preset angle, it restarts acquiring images and resets the number of escape attempts to zero.
[0234] If there is an obstacle in the range corresponding to the area to be driven in the image, the area to be driven is reduced; and the process proceeds to the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image.
[0235] One side of the area to be driven before shrinking is larger than the body width of the autonomous mobile robot, and the other side is the second preset distance; one side of the area to be driven after shrinking is smaller than the body by a preset width or equal to the body width, and the other side is the third preset distance, which is smaller than the second preset distance.
[0236] (2) Scene of being trapped in dense grass
[0237] The conditions for escape include: when the autonomous mobile robot is mowing in place, the actual speed of the cutter disc is less than or equal to the preset speed. During the process of driving and mowing, if the ratio between the actual speed of the cutter disc and the originally set expected speed is less than or equal to the preset value within a predetermined time, for example, if the ratio between the actual speed of the cutter disc and the expected speed is less than or equal to 55% within 3 seconds, it can be determined that the cutter disc is stuck and the machine is stuck, and the escape strategy needs to be initiated. When the machine is detected to have been stuck 25 times within 10 minutes, in order to avoid damage to the machine, it can be controlled to stop and the user can be reminded to assist the machine in handling the current situation. It is worth noting that since it takes some time for the cutter disc to start and reach the expected speed, the data detected within the first 5 seconds of the cutter disc startup needs to be ignored when determining whether the machine has experienced a cutter disc jam.
[0238] On the basis of the above-mentioned escape strategy applicable to multiple collision trapped scenarios, before controlling the autonomous mobile robot to rotate for the first time, the autonomous mobile robot is controlled to stop the cutter disc and retreat, and when retreating a preset distance, the autonomous mobile robot is controlled to start the cutter disc and start rotating; if the autonomous mobile robot is currently in a slipping trapped scenario, the autonomous mobile robot is controlled to rotate for the first time at a first speed, and adopts a different speed for each rotation, and the first speed is less than the preset speed.
[0239] If there are no obstacles within the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to drive out of the area to be driven to escape, including:
[0240] If there are no obstacles within the range corresponding to the area to be driven in the image and the vehicle is currently in a preset trapped scene, determining whether the rotation angle in the current rotation direction is greater than a third preset angle;
[0241] If the rotation angle is greater than the third preset angle, the autonomous mobile robot is controlled to drive out of the waiting area to escape.
[0242] If the rotation angle is less than or equal to the third preset angle, the autonomous mobile robot is controlled to rotate in the current rotation direction; and the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot is entered.
[0243] (3) Slip and get stuck
[0244] The conditions for escape include: the image acquired by the visual sensor remains unchanged for a predetermined period of time or the relationship between the expected value and the actual value of other sensors, such as the relationship between the IMU and the angular velocity, is used to detect whether the machine is slipping. For example, it is detected that the corner points in the image have not changed or have basically not changed within 6 seconds. For example, the difference between the expected cumulative value of the angular velocity within 2 seconds and the IMU yaw angle can be compared to see if it is greater than a preset threshold, or it can be detected whether the situation greater than the preset threshold is triggered 4 times within 10 seconds. If these conditions are met, it is confirmed that the machine is slipping, and the escape strategy corresponding to the slip can be initiated.
[0245] Based on the above escape strategy applicable to the scenario of being trapped by multiple collisions, the autonomous mobile robot is controlled to rotate for the first time at a first speed, and a different speed is used in each rotation, and the first speed is less than the preset speed.
[0246] If there are no obstacles within the range corresponding to the area to be driven in the image, the autonomous mobile robot is controlled to drive out of the area to be driven to escape, including:
[0247] If there are no obstacles within the range corresponding to the area to be driven in the image and the vehicle is currently in a preset trapped scene, determining whether the rotation angle in the current rotation direction is greater than a third preset angle;
[0248] If the rotation angle is greater than the third preset angle, the autonomous mobile robot is controlled to drive out of the waiting area to escape.
[0249] If the rotation angle is less than or equal to the third preset angle, the autonomous mobile robot is controlled to rotate in the current rotation direction; and the step of detecting whether there is an obstacle in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot is entered.
[0250] The specific implementation process of the escape strategy in the above scenario is described above.
[0251] The control method of the autonomous mobile robot of the present application detects whether the autonomous mobile robot meets the conditions for getting out of trouble; if the conditions for getting out of trouble are met, then the method detects whether there are obstacles in the range corresponding to the area to be driven in the image in front of the autonomous mobile robot; if there are no obstacles in the range corresponding to the area to be driven in the image, then the autonomous mobile robot is controlled to drive out of the area to be driven to get out of trouble. The present application detects whether there are obstacles in the area to be driven in front of the autonomous mobile robot based on images acquired by a visual sensor, and then controls the autonomous mobile robot to drive out of the area without obstacles to get out of trouble. Since visual recognition makes the driving direction of the autonomous mobile robot more targeted when getting out of trouble, it can effectively improve the autonomous mobile robot's efficiency in getting out of trouble in complex areas.
[0252] The present application also provides an autonomous mobile robot, the autonomous mobile robot comprising a visual sensor and a controller;
[0253] The visual sensor is used to obtain an image in front of the autonomous mobile robot;
[0254] The controller is used to implement the steps of any of the above-mentioned control methods for the autonomous mobile robot when executed.
[0255] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the direction determination method of the autonomous mobile robot according to various embodiments of the present invention described in the above-mentioned "Exemplary Direction Determination Method" section of this specification.
[0256] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0257] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.
[0258] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.
[0259] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.
[0260] Although only a few embodiments of the present invention are described and illustrated in this specification, those skilled in the art should readily envision other means or structures for performing the functions and / or obtaining the structures described herein, and each such change or modification is considered to be within the scope of the present invention.
Claims
1. A control method for an autonomous mobile robot, It is characterized in that include: Receiving near-field obstacle information sent by at least one sensor, wherein the at least one sensor is installed on the autonomous mobile robot; Based on the near-field obstacle information, generating an obstacle location map around the autonomous mobile robot; In response to receiving the near-field obstacle information multiple times within a limited time or space, an escape path is generated, wherein the generating the escape path includes: searching for a traversable direction based on the current position of the autonomous mobile robot and the obstacle position map.
2. The method according to claim 1, It is characterized in that The near-field obstacle information includes at least one of the following: collision information representing a contact collision of the autonomous mobile robot, lifting information representing the autonomous mobile robot being lifted, motor torque information representing a stall in the autonomous mobile robot's driving motor, and obstacle remote sensing information representing the existence of obstacles around the autonomous mobile robot.
3. The method according to any one of claims 1 to 2, It is characterized in that Based on the near-field obstacle information, an obstacle location map around the autonomous mobile robot is generated, including: A position on the map corresponding to the near-field obstacle information is marked as a near-field obstacle.
4. The method according to any one of claims 1 to 3, It is characterized in that The obstacle map is a local map of the periphery of the autonomous mobile robot, and the local map is updated as the autonomous mobile robot moves.
5. The method according to any one of claims 1 to 4, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a navigable direction includes: Expanding near-field obstacles in the obstacle location map according to an expansion coefficient, wherein the expansion coefficient is approximately equal to a width of the autonomous mobile robot; An area in the obstacle location map where the near-field obstacle does not exist is used as the traversable direction.
6. The method according to claim 5, It is characterized in that Expanding the near-field obstacles in the obstacle location map according to the expansion coefficient includes: Part of the near-field obstacles in the obstacle position map are expanded according to the expansion coefficient, and the part of the near-field obstacles includes obstacle remote sensing information representing obstacles existing around the autonomous mobile robot.
7. The method according to any one of claims 1 to 6, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a traversable direction includes: Obtain a search direction and a search range for near-field obstacles in the obstacle location map, and take a direction in which no near-field obstacles exist within the search range as the passable direction. The initial search direction is a direction opposite to the moving direction of the autonomous mobile robot, and the search range is an area determined by the search direction and beyond the area covered by the autonomous mobile robot on the map.
8. The method according to claim 7, It is characterized in that When the passable direction cannot be found in the obstacle location map along the initial search direction, the search direction is switched to a direction opposite to the initial search direction.
9. The method according to any one of claims 7 to 8, It is characterized in that The method further comprises: The search direction is kept unchanged until the autonomous mobile robot receives the near-field obstacle information again while traveling in the passable direction.
10. The method according to any one of claims 7 to 9, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a traversable direction includes: In the process of controlling the autonomous mobile robot to travel in the passable direction, when the near-field obstacle information is received within a duration period, the search direction is switched to a direction opposite to the initial search direction.
11. The method according to any one of claims 1 to 10, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a traversable direction includes: As the current position of the autonomous mobile robot is updated and the obstacle position map is updated, the traversable direction is adjusted accordingly.
12. The method according to any one of claims 1 to 11, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a traversable direction includes: When the autonomous mobile robot cannot search for the passable direction, the autonomous mobile robot is controlled to ignore the obstacle location map and adjust the posture, wherein the posture adjustment method includes at least one of the following: forward, backward, and rotation; A traversable direction is searched according to the current position of the autonomous mobile robot after the posture adjustment and the obstacle position map.
13. The method according to claim 12, It is characterized in that According to the current position of the autonomous mobile robot after the posture adjustment and the obstacle position map, searching for a traversable direction includes: A traversable direction is searched according to the current position of the autonomous mobile robot after adjusting the posture and the updated obstacle position map; wherein the obstacle position map is obtained after updating the obstacles in the obstacle position map according to preset rules, and the preset rules are that when the near-field obstacles recognized in real time by the obstacle remote sensing sensor installed on the autonomous mobile robot are inconsistent with the obstacles in the obstacle position map, the obstacle position map is updated according to the real-time recognition result of the obstacle remote sensing sensor.
14. The method according to any one of claims 1 to 13, It is characterized in that The moving angular velocity and the moving linear velocity are determined according to the angular difference between the moving direction of the autonomous mobile robot and the passable direction, and the autonomous mobile robot is controlled to move toward the passable direction according to the moving angular velocity and the moving linear velocity, wherein the moving angular velocity is positively correlated with the angular difference, and the moving linear velocity is negatively correlated with the angular difference.
15. The method according to any one of claims 1 to 14, It is characterized in that Based on the current position of the autonomous mobile robot and the obstacle position map, searching for a traversable direction includes: When controlling the autonomous mobile robot to move in the passable direction, determining whether the autonomous mobile robot can rotate according to an obstacle position map around the autonomous mobile robot; If it is rotatable, its rotation is controlled, and the passable direction is searched according to the position after rotation.
16. The method according to claim 15, It is characterized in that When determining whether it is rotatable, the initial rotation direction of the autonomous mobile robot when determining the rotation direction is consistent with the edge direction of the autonomous mobile robot.
17. The method according to claim 15 or 16, It is characterized in that Searching for the traversable direction according to the rotated position includes: In the process of controlling the rotation of the autonomous mobile robot, judging whether there is the near-field obstacle in front by obstacle remote sensing information collected in real time by an obstacle remote sensing sensor, wherein the obstacle remote sensing sensor is installed on the autonomous mobile robot, and the obstacle remote sensing sensor is configured to identify obstacle remote sensing information existing around the autonomous mobile robot; When it is determined that the near-field obstacle does not exist, the direction is used as the passable direction, and the autonomous mobile robot is controlled to move along the direction.
18. The method according to claim 17, It is characterized in that It is determined whether there is the near-field obstacle ahead through obstacle remote sensing information collected in real time by the obstacle remote sensing sensor and the obstacle position map.
19. The method according to any one of claims 1 to 18, It is characterized in that The method further comprises: When the autonomous mobile robot travels at least a first preset distance along the escape path, it is determined that the autonomous mobile robot has escaped successfully.
20. The method according to any one of claims 1 to 19, It is characterized in that The method further comprises: After the autonomous mobile robot successfully escapes, when it receives the near-field obstacle information multiple times within a limited time or space, the area covered in the obstacle map is marked as an obstacle, and a path is planned according to the re-marked map.
21. An autonomous mobile robot, It is characterized in that The autonomous mobile robot comprises: at least one sensor, the sensor being configured to acquire and transmit near-field obstacle information, the at least one sensor being mounted on the autonomous mobile robot; A controller is connected to the sensor signal, and is used to receive the near-field obstacle information, and generate an obstacle location map around the autonomous mobile robot based on the near-field obstacle information; in response to receiving the near-field obstacle information multiple times within a limited time or space, an escape path is generated, wherein the generating the escape path includes: searching for a passable direction based on the current position of the autonomous mobile robot and the obstacle location map.
22. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, which is loaded and executed by the processor to implement the control method of the autonomous mobile robot as described in claims 1-20.
Citation Information
Patent Citations
Moving robot obstacle avoiding method based on laser range finder
CN102541057A
Control system of robot
CN108803588A
Autonomous mobile robot relieving method fused with multi-sensor information
CN110244743A
Method and device for controlling robot to get out of trap and robot
CN111061215A
Escape method and device for sweeping robot, computer equipment and storage medium
CN111904346A
Cited By
Intelligent mower escape method and intelligent mower
CN120476830A
