Robot navigation obstacle avoidance method and device

Abstract

The invention relates to the technical field of robot obstacle avoidance, and particularly discloses a robot navigation obstacle avoidance method and device, and the method comprises the steps: accessing a camera network of a navigation scene, obtaining images of all cameras at regular time, recognizing the images, positioning the contour of an obstacle, and constructing a scene map according to the contour of the obstacle. Arranging the scene map according to the time sequence, and continuating the contour of the obstacle in the scene map at the latest moment; based on the extended scene map, determining a motion path of the robot as navigation information, and sending the navigation information to the robot; and determining the application frequency of an obstacle avoidance module of the robot based on the extended scene map, and sending the application frequency to the robot. The method is used in the field of robot control for limiting scenes, the obstacle position is pre-determined based on a camera network, a piece of large-probability navigation information which does not pass through the obstacle is determined according to the obstacle position, meanwhile, the application frequency of the obstacle avoidance module is adjusted according to the determined obstacle position, and resource consumption is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of robot obstacle avoidance, and in particular to a robot navigation obstacle avoidance method and device. Background Art

[0002] With the development of random intelligent technology, robots have gradually entered people’s production and life and are used to complete certain tasks; robots are very common in many occasions.

[0003] One of the basic functions of a robot is obstacle avoidance. The existing obstacle avoidance function is implemented by an obstacle avoidance module built into the robot. The obstacle avoidance module needs to be activated in real time and consumes a lot of resources. Under the condition of limited energy, the obstacle avoidance module consumes more resources and will affect the normal operation of the robot. How to reduce the resource consumption of the obstacle avoidance module is the technical problem that the technical solution of the present invention wants to solve. Summary of the invention

[0004] The object of the present invention is to provide a robot navigation obstacle avoidance method and device to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: A robot navigation obstacle avoidance method, the method comprising: Access the camera network of the navigation scene, obtain images from all cameras at regular intervals, identify the images, locate the outline of obstacles, and build a scene map based on the outline of obstacles; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; Arrange the scene map in chronological order, and extend the obstacle contour in the scene map at the latest moment; the extension process is the contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; Determine the robot's motion path based on the extended scene map and send it to the robot as navigation information; Determine the application frequency of the robot's obstacle avoidance module based on the extended scene map and send it to the robot; After receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information and controls the obstacle avoidance module based on the application frequency.

[0006] As a further solution of the present invention: the camera network connected to the navigation scene, regularly acquiring images from all cameras, identifying the images, locating obstacle contours, and constructing a scene map according to the obstacle contours includes the following steps: Access the camera network of the navigation scene and obtain images from all cameras at regular intervals; when the camera sends an image, it records the shooting position and shooting direction of the image; Read the scene outline of the navigation scene and construct a floor plan; Determine a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; Identify each image, locate the obstacle outline, and insert the obstacle outline into the mapping area; After each image is recognized, a plan view containing obstacle outlines is used as a scene map.

[0007] As a further solution of the present invention: the step of arranging the scene maps in chronological order and extending the obstacle contour in the scene map at the latest moment comprises: Read the time tags in the scene map and determine the time sequence according to the time tags; Arrange the scene maps based on the time sequence; Read the scene map at the latest moment, and for any obstacle outline in it, locate the same obstacle outline in the scene map at the previous moment; Determine changes based on the same obstacle profile located; Extending the obstacle outline in the scene map at the latest moment according to the changes; Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; is a preset minimum quantity; The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

[0008] As a further solution of the present invention: the step of determining the movement path of the robot based on the extended scene map and sending it to the robot as navigation information includes: Query the robot's movement starting point and movement end point in the extended scene map; Randomly generate a preset number of motion paths from the starting point to the end point; Calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information; The movement difficulty is determined by the path distance and the path resistance. The process of determining the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type.

[0009] As a further solution of the present invention: the step of randomly generating a preset number of motion paths from the motion starting point to the motion end point comprises: With the starting point of the movement as the center and the preset distance as the radius, a target selection circle is established; Eliminate the area corresponding to the obstacle outline in the target selection circle, and randomly select points on the edge of the target selection circle after eliminating the area as the target point; Take the target point as the new starting point of movement and select the next target point; The loop is executed until the target selection circle contains the end point of the movement; Connect the movement starting point, all target points and the movement end point to obtain the movement path.

[0010] As a further solution of the present invention: the step of determining the application frequency of the robot's obstacle avoidance module based on the extended scene map and sending it to the robot comprises: Calculate the shortest distance between each position in the navigation information and each obstacle outline; Select the minimum value among all the closest distances; The application frequency is determined by the inverse ratio of the minimum value and sent to the robot.

[0011] The technical solution of the present invention also provides a robot navigation obstacle avoidance device, the device comprising: A scene map construction module is used to access the camera network of the navigation scene, regularly obtain images from all cameras, identify the images, locate the outline of obstacles, and construct a scene map based on the outline of obstacles; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; The contour extension module is used to arrange the scene map in chronological order and extend the obstacle contour in the scene map at the latest moment; the extension process is the contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; A navigation information sending module is used to determine the movement path of the robot based on the extended scene map and send it to the robot as navigation information; An application frequency sending module, used to determine the application frequency of the robot's obstacle avoidance module based on the extended scene map, and send it to the robot; After receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information and controls the obstacle avoidance module based on the application frequency.

[0012] As a further solution of the present invention: the scene map construction module includes: An image acquisition unit is used to access the camera network of the navigation scene and regularly acquire images from all cameras; when the camera sends an image, it records the shooting position and shooting direction of the image; A plane map construction unit, used for reading the scene outline of the navigation scene and constructing a plane map; An area mapping unit, used to determine a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; A contour insertion unit, used to identify each image, locate the obstacle contour, and insert the obstacle contour into the mapping area; The map output unit is used to generate a plan view containing obstacle outlines as a scene map after each image is recognized.

[0013] As a further solution of the present invention: the contour extension module includes: A sequence determination unit, used for reading time tags in the scene map and determining the time sequence according to the time tags; A map arrangement unit, used for arranging scene maps based on the time sequence; A contour positioning unit is used to read the scene map at the latest moment, and for any obstacle contour therein, locate the same obstacle contour in the scene map at the previous moment; A change situation determination unit, used to determine the change situation according to the same obstacle outline located; An extension execution unit, used for extending the obstacle outline in the scene map at the latest moment according to the change; Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; is a preset minimum quantity; The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

[0014] As a further solution of the present invention: the navigation information sending module includes: A start and end point query unit is used to query the robot's movement start and end point in the extended scene map; A path generation unit, used for randomly generating a preset number of motion paths from a motion starting point to a motion end point; A selection unit is used to calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information; The movement difficulty is determined by the path distance and the path resistance. The process of determining the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention is used in the field of robot control in a limited scene, predetermines the position of obstacles based on a camera network, and determines a navigation information that is unlikely to pass through the obstacle according to the position of the obstacle. At the same time, the application frequency of the obstacle avoidance module is adjusted according to the determined obstacle position, thereby effectively reducing resource consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0017] Figure 1 Flow chart of the robot navigation and obstacle avoidance method.

[0018] Figure 2 This is the first sub-flow chart of the robot navigation and obstacle avoidance method.

[0019] Figure 3 This is the second sub-process flowchart of the robot navigation and obstacle avoidance method.

[0020] Figure 4 This is the third sub-process flowchart of the robot navigation and obstacle avoidance method.

[0021] Figure 5 This is the fourth sub-process flowchart of the robot navigation and obstacle avoidance method.

[0022] Figure 6 This is a structural block diagram of the robot navigation and obstacle avoidance system. DETAILED DESCRIPTION

[0023] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0024] Figure 1 : is a flowchart of a robot navigation and obstacle avoidance method. In an embodiment of the present invention, a robot navigation and obstacle avoidance method comprises: Step S100: accessing the camera network of the navigation scene, periodically acquiring images from all cameras, identifying the images, locating obstacle outlines, and constructing a scene map according to the obstacle outlines; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; The navigation scene is the working scene of the robot. In the existing working scenes of the robot, multiple video acquisition devices are generally installed, collectively referred to as a camera network. The camera network is a collection of cameras. The images of all cameras are acquired at a preset frequency and recognized to locate obstacles and determine their outlines. All obstacle outlines are counted to build a scene map. The accuracy of the scene map does not need to be too high, it only needs to be able to display the obstacle outlines. In addition, since the scene map is acquired based on images containing time, the scene map also contains time information. Since the process of acquiring images from all cameras is timed, the time differences of all images are very small. There will only be some time differences caused by the instruction transmission process. The average of all times is calculated as a unified result.

[0025] Step S200: Arrange the scene map in chronological order, and extend the obstacle contour in the scene map at the latest moment; the extension process is a contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; The scene maps are arranged in chronological order. Each scene map contains obstacle outlines. The obstacles include still objects and animals. Regardless of the type of obstacle, its shape will not change significantly in a short period of time. Therefore, in the arranged scene maps, the outline corresponding to the same obstacle can be easily located, and then its change can be determined; each obstacle outline can be extended according to the change.

[0026] Step S300: determining the movement path of the robot based on the extended scene map, and sending the movement path to the robot as navigation information; The extension process is equivalent to expanding the obstacle outline in the scene map to obtain multiple obstacle areas, and determining the robot's motion path based on the obstacle areas, which are sent to the robot as navigation information. This process is actually based on the camera network to determine a motion path that will hardly touch obstacles from a macro perspective.

[0027] Step S400: determining the application frequency of the robot's obstacle avoidance module based on the extended scene map, and sending it to the robot; On the basis of determining the motion path, the application frequency of the robot's obstacle avoidance module is determined based on the extended scene map. The ultimate goal of this application is to minimize the application frequency of the robot's obstacle avoidance module. Since the determined motion path itself will not pass through too many obstacles, on this basis, the application frequency of the obstacle avoidance module can be adjusted according to the actual distance between the robot and the contours of each obstacle. The application frequency of the obstacle avoidance module will pre-set a minimum value to prevent emergencies and send the application frequency to the robot.

[0028] On the robot side, after receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information, mainly controlling the movement direction. The movement speed is the speed pre-set by the robot, which is generally a uniform movement speed. The obstacle avoidance module is controlled based on the application frequency. The application frequency indicates how often the robot-based obstacle avoidance behavior is performed. The obstacle avoidance module is generally some distance sensors used to detect whether there are obstacles nearby.

[0029] Figure 2 The first sub-flow diagram of the robot navigation and obstacle avoidance method is shown in FIG. 1 . The camera network connected to the navigation scene, regularly acquiring images from all cameras, identifying the images, locating the obstacle outline, and constructing a scene map according to the obstacle outline include the following steps: Step S101: accessing the camera network of the navigation scene, and acquiring images from all cameras at regular intervals; when the cameras send images, the shooting positions and directions of the images are recorded; Step S102: reading the scene outline of the navigation scene and constructing a plan view; Step S103: determining a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; Step S104: Identify each image, locate the obstacle outline, and insert the obstacle outline into the mapping area; Step S105: After each image is identified, the plane image containing the obstacle outline is used as the scene map.

[0030] In an example of the technical solution of the present invention, the process of constructing the scene map is specifically limited, the camera network of the navigation scene is connected, and images of all cameras are acquired at regular intervals. When the camera acquires the image, the shooting position and shooting direction are recorded. When the camera uploads the image, the recorded shooting position and shooting direction are uploaded together.

[0031] Furthermore, the boundary of the navigation scene is read, called the scene outline (usually a bird's-eye view), and a plan view is constructed. According to the shooting position and shooting direction, it can be determined which area in the plan view the shooting picture corresponds to (the intersection of the acquisition range in the plan view), which is called the mapping area. Then, each image is identified, the obstacle outline is located, and the obstacle outline is inserted into the mapping area. The located obstacle outline is a collection of outline point positions. After each image is identified, the plan view containing the obstacle outline is used as the scene map.

[0032] Figure 3 The second sub-flow chart of the robot navigation obstacle avoidance method is as follows. The step of arranging the scene map in chronological order and extending the obstacle outline in the scene map at the latest moment includes: Step S201: Read the time tags in the scene map and determine the time sequence according to the time tags; Step S202: Arranging scene maps based on the time sequence; Step S203: reading the scene map at the latest moment, and locating the same obstacle outline in the scene map at the previous moment for any obstacle outline therein; Step S204: determining the change according to the same obstacle outline located; Step S205: Extend the obstacle outline in the scene map at the latest moment according to the change.

[0033] In an example of the technical solution of the present invention, the extension process of the obstacle contour is described in detail, the time tag in the scene map is read, the time sequence is determined according to the time tag, the scene map is arranged based on the time sequence, the scene map at the latest moment is read, and for any obstacle contour therein, the same obstacle contour is located in the scene map at the previous moment according to the shape of the contour. The same obstacle refers to the obstacle contour corresponding to the same obstacle in different images. By comparing the same obstacle contour at different moments, the change can be determined, and the obstacle contour in the scene map at the latest moment is extended according to the change.

[0034] Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; The preset minimum quantity.

[0035] As for step S203, its purpose is to select some scene images for calculating the change. The number of selected images is dynamic. This is a unique solution of the present application. The principle is to continuously acquire scene images from the scene image at the latest moment, and compare the difference between the same obstacle contour in the two scene images. If the difference is large, continue to acquire scene images and analyze them at the same time. If the difference is not large, output the number of scene images that have been acquired as the final number; wherein, The minimum number of scene images to be acquired is limited. The value can be five or ten.

[0036] The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

[0037] Furthermore, regarding the extension process, for any obstacle contour, after processing, it has a corresponding number of scene maps. The scene map at the earliest moment and the scene map at the latest moment are selected from the acquired scene maps, and the centroid is determined based on the same centroid calculation scheme (conventional mathematical means). The centroids are connected to obtain a vector, called the change vector; the obstacle contour is extended on the change vector; the extension direction is the same as the direction of the change vector, and the extension length is proportional to the modulus of the change vector; in addition, regarding the extension width, a direction perpendicular to the extension direction is generally determined as the width direction, and the maximum value of the obstacle contour is queried in the width direction as the extension width.

[0038] Figure 4 The third sub-flow chart of the robot navigation and obstacle avoidance method is shown in FIG. 1 , wherein the step of determining the movement path of the robot based on the extended scene map and sending the navigation information to the robot as the navigation information includes: Step S301: querying the robot's movement starting point and movement end point in the extended scene map; Step S302: randomly generating a preset number of motion paths from the motion starting point to the motion end point; Step S303: Calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information.

[0039] In an example of the technical solution of the present invention, the process of generating and selecting the motion path is described in detail. First, no matter which robot, the starting point and end point of its working process are known, and the robot's motion starting point and motion end point are queried in the extended scene map; then, a preset number of motion paths pointing from the motion starting point to the motion end point are randomly generated. Since the number of generated motion paths is not unique, a selection process is required to calculate the motion difficulty of each motion path, and the final motion path is selected according to the motion difficulty, and sent to the robot as navigation information.

[0040] Specifically, the movement difficulty is a special term of the present invention, and the movement difficulty is determined by the path distance and the path resistance. The determination process of the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type; the longer the path distance, the higher the movement difficulty, and the larger the path resistance, the higher the movement difficulty. Regarding the specific mapping relationship, a direct proportional function, an exponential function or a logarithmic function is generally used, and they have different derivative characteristics. Of course, some combinations can also be performed, as long as the correlation that the longer the path distance, the higher the movement difficulty, the larger the path resistance, the higher the movement difficulty is satisfied.

[0041] Furthermore, the step of randomly generating a preset number of motion paths from the motion starting point to the motion end point includes: With the starting point of the movement as the center and the preset distance as the radius, a target selection circle is established; Eliminate the area corresponding to the obstacle outline in the target selection circle, and randomly select points on the edge of the target selection circle after eliminating the area as the target point; Take the target point as the new starting point of movement and select the next target point; The loop is executed until the target selection circle contains the end point of the movement; Connect the movement starting point, all target points and the movement end point to obtain the movement path.

[0042] The above content specifically describes the generation process of any motion path, which is a process of continuously selecting the next point, which is equivalent to continuously selecting waypoints from the starting point of the motion until the end point of the motion is reached; specifically, each time the next waypoint is selected, the distance between the next waypoint and the current position should be fixed, that is, the radius in the above content; with the starting point of the motion as the center and the preset distance as the radius, a target selection circle is established, and the area corresponding to the obstacle outline is eliminated in the target selection circle, so that obstacles can be avoided, and points are randomly selected on the edge of the target selection circle after the area is eliminated as target points, that is, waypoints; the target point is used as the new starting point of the motion, and the next target point is selected, and the cycle is executed until the target selection circle contains the end point of the motion, and finally, the starting point of the motion, all target points and the end point of the motion are connected to obtain the motion path.

[0043] The above motion path generation process is performed multiple times to obtain a preset number of motion paths.

[0044] Figure 5 The fourth sub-flow chart of the robot navigation obstacle avoidance method, wherein the step of determining the application frequency of the robot's obstacle avoidance module based on the extended scene map and sending the map to the robot includes: Step S401: Calculate the shortest distance between each position in the navigation information and each obstacle outline; Step S402: Select the minimum value among all the closest distances; Step S403: Determine the application frequency according to the inverse ratio of the minimum value and send it to the robot.

[0045] In the technical solution of the present invention, the navigation information is the motion path finally selected, and the closest distance between each position in the navigation information and the outline of each obstacle is calculated. The minimum value is selected among all the closest distances. The minimum value indicates how far the obstacle closest to the current position is. The application frequency is determined according to the inverse proportion of the minimum value and sent to the robot; that is, the larger the minimum value, the farther away from the obstacle, and the smaller the application frequency.

[0046] Figure 6 : is a structural block diagram of a robot navigation and obstacle avoidance system. In an embodiment of the present invention, a robot navigation and obstacle avoidance device, the device 10 includes: The scene map construction module 11 is used to access the camera network of the navigation scene, regularly obtain images from all cameras, identify the images, locate the outline of obstacles, and construct a scene map according to the outline of obstacles; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; The contour extension module 12 is used to arrange the scene map in chronological order and extend the obstacle contour in the scene map at the latest moment; the extension process is a contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; A navigation information sending module 13 is used to determine the movement path of the robot based on the extended scene map and send it to the robot as navigation information; An application frequency sending module 14 is used to determine the application frequency of the robot's obstacle avoidance module based on the extended scene map and send it to the robot; After receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information and controls the obstacle avoidance module based on the application frequency.

[0047] Furthermore, the scene map construction module 11 includes: An image acquisition unit is used to access the camera network of the navigation scene and regularly acquire images from all cameras; when the camera sends an image, it records the shooting position and shooting direction of the image; A plane map construction unit, used for reading the scene outline of the navigation scene and constructing a plane map; An area mapping unit, used to determine a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; A contour insertion unit, used to identify each image, locate the obstacle contour, and insert the obstacle contour into the mapping area; The map output unit is used to generate a plan view containing obstacle outlines as a scene map after each image is recognized.

[0048] Specifically, the contour extension module 12 includes: A sequence determination unit, used for reading time tags in the scene map and determining the time sequence according to the time tags; A map arrangement unit, used for arranging scene maps based on the time sequence; A contour positioning unit is used to read the scene map at the latest moment, and for any obstacle contour therein, locate the same obstacle contour in the scene map at the previous moment; A change situation determination unit, used to determine the change situation according to the same obstacle outline located; An extension execution unit, used for extending the obstacle outline in the scene map at the latest moment according to the change; Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; is a preset minimum quantity; The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

[0049] Furthermore, the navigation information sending module 13 includes: A start and end point query unit is used to query the robot's movement start and end point in the extended scene map; A path generation unit, used for randomly generating a preset number of motion paths from a motion starting point to a motion end point; A selection unit is used to calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information; The movement difficulty is determined by the path distance and the path resistance. The process of determining the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type.

[0050] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A robot navigation obstacle avoidance method, characterized in that: The method comprises: Access the camera network of the navigation scene, obtain images from all cameras at regular intervals, identify the images, locate the outline of obstacles, and build a scene map based on the outline of obstacles; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; Arrange the scene map in chronological order, and extend the obstacle contour in the scene map at the latest moment; the extension process is the contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; Determine the robot's motion path based on the extended scene map and send it to the robot as navigation information; Determine the application frequency of the robot's obstacle avoidance module based on the extended scene map and send it to the robot; After receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information and controls the obstacle avoidance module based on the application frequency.

2. The robot navigation obstacle avoidance method according to claim 1, characterized in that: The steps of accessing the camera network of the navigation scene, periodically acquiring images from all cameras, identifying the images, locating obstacle contours, and constructing a scene map according to the obstacle contours include: Access the camera network of the navigation scene and obtain images from all cameras at regular intervals; when the camera sends an image, it records the shooting position and shooting direction of the image; Read the scene outline of the navigation scene and construct a floor plan; Determine a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; Identify each image, locate the obstacle outline, and insert the obstacle outline into the mapping area; After each image is recognized, a plan view containing obstacle outlines is used as a scene map.

3. The robot navigation obstacle avoidance method according to claim 1, characterized in that: The step of arranging the scene maps in time order and extending the obstacle contour in the scene map at the latest moment comprises: Read the time tags in the scene map and determine the time sequence according to the time tags; Arrange the scene maps based on the time sequence; Read the scene map at the latest moment, and for any obstacle outline in it, locate the same obstacle outline in the scene map at the previous moment; Determine changes based on the same obstacle profile located; Extending the obstacle outline in the scene map at the latest moment according to the changes; Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; is a preset minimum quantity; The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

4. The robot navigation obstacle avoidance method according to claim 1, characterized in that: The step of determining the movement path of the robot based on the extended scene map and sending the navigation information to the robot comprises: Query the robot's movement starting point and movement end point in the extended scene map; Randomly generate a preset number of motion paths from the starting point to the end point; Calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information; The movement difficulty is determined by the path distance and the path resistance. The process of determining the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type.

5. The robot navigation obstacle avoidance method according to claim 4, characterized in that: The step of randomly generating a preset number of motion paths from the motion starting point to the motion end point comprises: With the starting point of the movement as the center and the preset distance as the radius, a target selection circle is established; Eliminate the area corresponding to the obstacle outline in the target selection circle, and randomly select points on the edge of the target selection circle after eliminating the area as the target point; Take the target point as the new starting point of movement and select the next target point; The loop is executed until the target selection circle contains the end point of the movement; Connect the movement starting point, all target points and the movement end point to obtain the movement path.

6. The robot navigation obstacle avoidance method according to claim 1, characterized in that: The step of determining the application frequency of the robot's obstacle avoidance module based on the extended scene map and sending the map to the robot comprises: Calculate the shortest distance between each position in the navigation information and each obstacle outline; Select the minimum value among all the closest distances; The application frequency is determined by the inverse ratio of the minimum value and sent to the robot.

7. A robot navigation and obstacle avoidance device, characterized in that: The device comprises: A scene map construction module is used to access the camera network of the navigation scene, regularly obtain images from all cameras, identify the images, locate the outline of obstacles, and construct a scene map based on the outline of obstacles; the scene map contains a time tag, and the time tag of the scene map is the average of the time tags of the acquired images; The contour extension module is used to arrange the scene map in chronological order and extend the obstacle contour in the scene map at the latest moment; the extension process is the contour expansion process, and the direction and degree of contour expansion are determined by the change of the obstacle contour; A navigation information sending module is used to determine the movement path of the robot based on the extended scene map and send it to the robot as navigation information; An application frequency sending module, used to determine the application frequency of the robot's obstacle avoidance module based on the extended scene map, and send it to the robot; After receiving the navigation information and application frequency, the robot controls the movement process based on the navigation information and controls the obstacle avoidance module based on the application frequency.

8. The robot navigation and obstacle avoidance device according to claim 7, characterized in that: The scene map construction module includes: An image acquisition unit is used to access the camera network of the navigation scene and regularly acquire images from all cameras; when the camera sends an image, it records the shooting position and shooting direction of the image; A plane map construction unit, used for reading the scene outline of the navigation scene and constructing a plane map; An area mapping unit, used to determine a mapping area of ​​the image in the plane map according to the shooting position and the shooting direction; A contour insertion unit, used to identify each image, locate the obstacle contour, and insert the obstacle contour into the mapping area; The map output unit is used to generate a plan view containing obstacle outlines as a scene map after each image is recognized.

9. The robot navigation and obstacle avoidance device according to claim 7, characterized in that: The contour extension module includes: A sequence determination unit, used for reading time tags in the scene map and determining the time sequence according to the time tags; A map arrangement unit, used for arranging scene maps based on the time sequence; A contour positioning unit is used to read the scene map at the latest moment, and for any obstacle contour therein, locate the same obstacle contour in the scene map at the previous moment; A change situation determination unit, used to determine the change situation according to the same obstacle outline located; An extension execution unit, used for extending the obstacle outline in the scene map at the latest moment according to the change; Among them, the number of scene maps at the previous moment includes: ; In the formula, Indicates the number of scene maps, After arranging the scene maps in the time order, the first The area of ​​the image, After arranging the scene maps in the time order, the first The area of ​​the image; is the preset threshold; is a preset minimum quantity; The process of determining the changes is: Among the acquired scene maps, select the scene map at the earliest moment and the scene map at the latest moment; Determine the centroid of the same obstacle outline in the scene map at the earliest moment and the scene map at the latest moment, connect the centroids, and obtain a change vector; the modulus of the change vector is the distance of the centroid, and the direction of the change vector is from the centroid at the previous moment to the centroid at the next moment; The extension direction and the extension length are determined based on the change vector, and the obstacle contour is extended based on the extension direction and the extension length; the extension width adopts the width of the obstacle contour.

10. The robot navigation and obstacle avoidance device according to claim 7, characterized in that: The navigation information sending module comprises: A start and end point query unit is used to query the robot's movement start and end point in the extended scene map; A path generation unit, used for randomly generating a preset number of motion paths from a motion starting point to a motion end point; A selection unit is used to calculate the movement difficulty of each movement path, select the final movement path according to the movement difficulty, and send it to the robot as navigation information; The movement difficulty is determined by the path distance and the path resistance. The process of determining the path resistance is to read the ground type on the movement path and query the path resistance in a preset resistance table according to the ground type.

Images