WO2014075598A1

WO2014075598A1 - Mobile robot separating visual positioning and navigation method and positioning and navigation system thereof

Info

Publication number: WO2014075598A1
Application number: PCT/CN2013/086908
Authority: WO
Inventors: 朱绍明
Original assignee: Zhu Shaoming
Priority date: 2012-11-13
Filing date: 2013-11-12
Publication date: 2014-05-22
Also published as: CN102929280B; CN102929280A

Abstract

A mobile robot separating visual positioning and navigation method comprises that corresponding site images are collected in real time above a target work site; the target work site is recognized in the site images; the recognition of the target work site comprises the recognition of target work site boundaries and barriers; according to a preset working mode, a corresponding working path is generated on the recognized target work site; according to the working path, a mobile robot is instructed to work. The present invention is high in efficiency, accurate in positioning and navigation and low in cost, and simultaneously can efficiently achieve a path patrol type application and a traversal scan type application.

Description

Mobile robot separated visual positioning navigation method and positioning navigation system thereof

The invention belongs to the technical field of mobile robots, and in particular relates to a mobile robot separated visual positioning and navigation method and a positioning navigation system thereof. Background technique

Currently, some workplace work can or has been done automatically using mobile robot systems, such as lawn mowers, agricultural uses, door snow cleaning, golf training, ball cleaning, indoor cleaning, factory material handling, and more.

However, the positioning and navigation method of the traditional mobile robot system has the disadvantages of low flexibility, low efficiency, inaccurate positioning navigation or high cost. Some of the above problems exist in the traditional mobile robot systems: (1) Low-voltage wires are laid in the field, and the conductive wires are detected by the sensors on the robot for boundary recognition and positioning and guiding, but the method is troublesome in wiring and complicated in operation. It requires cumbersome wiring or only low-efficiency random scanning. (2) By laser scanning the triangulation, the site is digitized by this method, and then the robot is manually remotely operated to identify the boundary and obstacles. This method is troublesome to install and reduces overall efficiency. (3) The laser rangefinder scanning is used to establish a site map by laser ranging scanning, which is generally used for point-to-point operation and requires professional installation. Outdoors can be used with low-resolution GPS initial positioning assistance, and the cost is high. (4) Through color tracking, the running route is marked with a special color. The camera or sensor located on the robot navigates by recognizing the color. This method is similar to the low-pressure guiding method. When the path is changed, the color code needs to be re-marked, which is inefficient. (5) The sensor automatically detects the boundary method. The sensor automatically detects the boundary and obstacles, runs straight in the middle of the field, and randomly moves at an angle to continue running when the boundary is encountered. There is no overall concept, it may work repeatedly, the accuracy is not high, and the efficiency is low. Summary of the invention

Based on this, the present invention is directed to the above technical problem, and provides a mobile robot separate view. Sense positioning navigation method and its positioning navigation system.

The invention adopts the following technical solutions:

A mobile robot separate visual positioning and navigation method, comprising:

Collect corresponding site images in real time above the target work site;

Identifying a target work site in the site image, including identifying a target work site boundary and an obstacle;

According to the preset working mode, a corresponding working path is generated on the identified target work site; according to the working path, the mobile robot is instructed to work.

In one of the embodiments, the step of identifying the target worksite in the scene image further comprises scaling the target worksite to generate a mapping between the pixel points and the actual points.

In one of the embodiments, the preset working mode includes a patrol mode and an traversal mode. In one embodiment, when the preset working mode is the patrol mode, the step of generating a corresponding working path on the identified target work site includes:

Determine the working point and the working path according to the preset;

Converting the working point and the working path into corresponding pixel point coordinate sets;

Set the corresponding work action command at the working point according to the preset.

In one embodiment, when the preset working mode is the traversing mode, the step of generating a corresponding working path on the identified target work site includes:

Traversing the image of the site to generate a zigzag working path;

The zigzag working path is converted into a corresponding set of pixel point coordinates.

In one embodiment, the instructing the robot working step according to the working path comprises:

Reading the working path one by one, and instructing the mobile robot to run according to the pixel point coordinate set of the working path, and the corresponding working action instruction instructs the mobile robot to perform corresponding working actions at the working point;

The deviation value of the mobile robot from the current working path is calculated in real time, and if the deviation value is greater than the preset maximum deviation value, the mobile robot is instructed to return to the current working path.

In one embodiment, a mobile robot separate visual positioning navigation system includes a camera, a control device for acquiring a corresponding scene image in real time, and a mobile robot that is instructed by the control device to operate, the camera is disposed above the target work site, and the control device is connected to the camera and the mobile robot signal;

The control device includes:

a target work site identification unit, configured to identify a target work site in the site image, including identifying a target work site boundary and an obstacle;

a path generating unit, configured to generate a corresponding working path on the identified target work site according to the preset working mode;

And a control unit, configured to instruct the mobile robot to work according to the working path.

In one embodiment, the control device further includes a scaling unit for scaling the target work site to generate a mapping between the pixel points and the actual points.

In one embodiment, the path generating unit includes:

The patrol path generation module is configured to determine the working point and the working path according to the preset, and convert the working point and the working path into corresponding pixel point coordinate sets, and then set corresponding work action instructions according to the preset at the working point.

The traversing path generating module is configured to traverse the scanned scene image, generate a zigzag working path, and convert the zigzag working path into a corresponding pixel point coordinate set.

In one embodiment, the control unit comprises:

The instruction module is configured to read the working path one by one, and instruct the mobile robot to run according to the pixel point coordinate set of the working path, and at the working point, the corresponding working action instruction instructs the mobile robot to perform a corresponding working action;

The judging module is configured to calculate the deviation value of the mobile robot from the current working path in real time, and if the deviation value is greater than the preset maximum deviation value, instruct the mobile robot to return to the current working path.

The invention has high efficiency, accurate positioning and navigation, convenient line changing, and low cost, and can effectively implement both path patrol and traversal scanning. DRAWINGS

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments:

1 is a flow chart of a mobile robot separate visual positioning and navigation method according to the present invention Figure.

2 is a flow chart of a process for generating a working path when the working mode is a patrol mode; FIG. 3 is a flow chart of a process for generating a working path when the working mode is a traversal mode; FIG. 4 is an embodiment of the present invention. Flow chart of step 140;

FIG. 5 is a schematic diagram of a mobile robot separate visual positioning navigation system according to the present invention; FIG.

6 is a schematic diagram of a working path of a patrol mode according to the present invention;

7 is a schematic diagram of a working path of an ergodic mode according to the present invention;

Fig. 8 is a schematic structural view of a control device of the present invention. detailed description

As shown in FIG. 1 , a mobile robot separate visual positioning and navigation method includes: S 110, real-time collecting corresponding scene images on a target work site;

S120. Identify a target work site in the site image, including identifying a target work site boundary and an obstacle;

Specifically, the range of the mobile robot to be operated, that is, the boundary, the obstacle, and the like can be set by automatic recognition, human-computer interaction recognition, or manual recognition. Make different treatments. For example, when used in lawn applications, it can be set as a pond, sand pit, rockery, tree, fence, etc. When it is a pond or bunker, you can use a sounding sensor or redundant visual positioning (ie leave a margin) to ensure that the robot does not fall. Rockeries, trees, and fences can be identified with collision sensors without the need for redundant visual positioning.

If there are inaccessible obstacles such as ponds or bunkers in the site, then the target work site must be scaled, and then the mobile robot can be monitored manually in the work site. Of course, it can also be remotely moved by the remote control. Robot.

Since the mobile robot has a physical size, a mapping between pixel points and actual points can be established during its operation. 5厘米。 The size of the actual corresponding size of each pixel is 0. 5cm. machine S 1 30. Generate a corresponding working path on the identified target work site according to a preset working mode;

The preset working modes include a patrol mode and an traversal mode.

Specifically, as shown in FIG. 2, when the working mode is the patrol mode, the working path is generated as follows:

S 1 3 la, determining the working point and the working path according to the preset; wherein the working point is usually preset, generally there are multiple working points, after the working point is determined, the working path is between the working point and the working point, and the working path can be Indicates the path, of course, you can also use the curve.

S I 32a, converting the working point and the working path into corresponding pixel point coordinate sets;

S 1 33a, according to the preset, set the corresponding work action instruction at the working point. As shown in Figure 6, different work points a, b, c need to set actions such as workpiece picking and placement.

It can be understood that if the robot is operated in this mode, it is not necessary to calibrate the work site in the step S120, and to avoid the inaccessible obstacle when determining the working path.

When the working mode is traversal mode, the working path is generated as follows:

S 1 31b, traversing the scanned field image to generate a zigzag working path;

Specifically, as shown in Fig. 3, one of the longest sides of the scene image is selected, or an edge is manually selected, along which a series of parallel lines are generated, the distance between the parallel lines being the working diameter of the mobile robot. These parallel lines intersect the boundary and the obstacle boundary to divide it into shorter segments. Then start with a parallel line, see Figure 7, which starts from the leftmost flat line segment. When the line segment is scanned, it looks for the nearest parallel line and scans it. This cycle, scanning the workplace. When a line is scanned, its adjacent sites have been scanned, and there are still areas that have not been scanned elsewhere. In this case, you need to cross the scanned area and continue scanning without scanning. When crossing an area that has already been scanned, the dotted line in the figure shows this path. After reaching the new area, scan again as described above. This cycle until the scan of all areas of the work site is completed.

S 1 32b, converting the zigzag working path into a corresponding pixel point coordinate set.

The above two path generation modes are converted into pixels, and then positioned and navigated according to the image. Of course, it can also be converted into actual geometric space, and positioning navigation in the geometric space.

S140, according to the working path, instructing the mobile robot to work, as shown in FIG. The specific process is as follows:

5141. Read the working path one by one, and instruct the mobile robot to run according to the pixel point coordinate set of the working path, and at the working point, the corresponding working action instruction instructs the mobile robot to perform the corresponding working action;

5142. During the working process, continue to collect the image of the site, and calculate the deviation value of the mobile robot from the current working path in real time. If the deviation value is greater than the preset maximum deviation value, the mobile robot is instructed to return to the current working path. After completing a path, read the next path and complete the above method. Until all instructions are completed.

For the traversal mode, it is also possible to run directly without generating a path in advance. The mobile robot operates in one direction, and its internal mechanism ensures that it runs straight in one direction. When it detects that it has reached the boundary or obstacle, it will give it an instruction, it will automatically turn, move the specified distance and then run in the opposite direction. This repeated operation, also works in a zigzag route, when the area that can be reached by continuous operation has been scanned, and there is still no scan elsewhere. At this time, the mobile robot judges that the command issued according to the computer image is empty to continue scanning without scanning the area. This also allows scanning of the entire workplace. It will be appreciated that if the robot is operating in this traversal mode, it is also possible to calibrate the work site in step S120.

When the power of the mobile robot is insufficient, the mobile robot will automatically return to the charging station to charge, and then return to the original position to continue working after charging. When the robot completes the work of the entire site, it will automatically return to the charging station for standby until the next task.

As shown in FIG. 5, the present invention also relates to a mobile robot separate visual positioning navigation system, comprising a camera 110 for real-time acquisition of an image of a corresponding work site 2, a control device 120, and instructions for issuing according to the control device 120. The mobile robot 1 30 is operated, the camera 110 is disposed above the target work site 2, and the control device 120 is signal-connected to the camera 110 and the mobile robot 130.

The robot 1 30 will work in the work site 2, which includes different shapes or different types of obstacles 4 . The camera 5 is fixed above the work site 2 and communicates with the control device 120 via wireless or wired.

Wherein, the camera 110 is based on the size of the work site 2 and the work of the mobile robot 1 30 Choose the precision. For example, if a working position of lQmxlQm is required, the position accuracy of the last mobile robot 130 is 1cm. If only one camera is used to complete the task, the required pixels are:

( 10m / (0.01m) *Si ) * ( 10m / (0.01m) * Si )

(Si-safety factor, determined by factors such as image quality and environment, greater than or equal to 2).

If the safety factor is 2, you need to select a lens with 2000 * 2000 pixels.

Locating the selected camera 110 to the appropriate position of the work site 2, so that these cameras

The field of view of 110 covers the field as evenly as possible, i.e., the same field area occupies the same number of pixels. However, since the camera is not located directly above the lawn and generally has an angle with the site, it is not possible to occupy the same number of pixels in the same site area.

The user can observe the captured image to adjust and confirm the optimal position of the camera.

It can be understood that if there are many obstacles, large areas or slender shapes in the workplace, multi-camera monitoring is required. When images are captured by multiple cameras in some places, the image with the slowest pixel change rate, that is, the image with the highest resolution in the area, is used to position the navigation.

The lens of the camera 110 can also be automatically telescopic, which can be used for overall monitoring, as well as local tracking of the robot with high-precision positioning and navigation.

Control device 120 can be independent of mobile robot 130 or located in a mobile robot

Among 130. When the control device 120 is independent of the mobile robot 130, communication with the mobile robot 130 is typically via a wireless network.

As shown in FIG. 8, specifically, the control device 120 includes:

a target worksite identification unit 121, configured to identify a target work site in the site image, including identifying a target worksite boundary 3 and an obstacle 4;

The path generating unit 122 is configured to generate a corresponding working path 5 on the identified target working field according to the preset working mode. The path generating unit 122 includes:

The patrol path generating module 122a is configured to determine the working point 6 and the working path 5 according to the preset, and convert the working point 6 and the working path 5 into corresponding pixel point coordinate sets, so as to set corresponding work at the working point 6 according to the preset. Action instruction.

The traversing path generating module 122b is configured to traverse the scanning site image, generate a working path 5, and convert the zigzag working path into a corresponding pixel point coordinate set, wherein the working path 5 is a word Shape.

The control unit 12 3 is configured to instruct the mobile robot 1 30 to operate according to the working path 5, and the control unit 12 3 includes:

The instruction module 12 3a is configured to read the working path 5 one by one, and instruct the mobile robot to run according to the pixel point coordinate set of the working path, and the corresponding working action instruction instructs the mobile robot to perform a corresponding working action at the working point;

The judging module 12 3b is configured to calculate the deviation value of the mobile robot from the current working path in real time, and if the deviation value is greater than the preset maximum deviation value, instruct the mobile robot to return to the current working path.

The control device 120 also includes a scaling unit 124 for scaling the target work site to generate a mapping between the pixel points and the actual points.

The mobile robot 1 30 is composed of a communication system, an operating system, a working system, and a sensing system.

Mobile robots 1 30 can be equipped with other sensors, such as sensors for collision, sounding, dumping, and removal.

The mobile robot 1 30 surface can be used in special color shapes or special color LED lights for easy identification from the working environment. It is usually equipped with an automatic charging station. When the robot is running low, the robot will automatically return to the charging station for charging.

The system can also be configured with an external device such as a remote controller 140 for presetting preset operating point and working mode related parameters, and can send work instructions to the control device 120 for manual operation.

It is to be understood by those skilled in the art that the above-described embodiments are only intended to illustrate the invention, and are not intended to limit the invention, as long as it is within the spirit of the invention, Variations and modifications are intended to fall within the scope of the appended claims. It is to be understood by those skilled in the art that the above-described embodiments are only intended to illustrate the invention, and are not intended to limit the invention, as long as it is within the spirit of the invention, Variations and modifications are intended to fall within the scope of the appended claims.

Claims

claims

1. A separate visual positioning and navigation method for a mobile robot, which is characterized by: collecting corresponding site images in real time above the target working site;

Identify the target work site in the site image, including identifying the boundaries and obstacles of the target work site;

According to the preset working mode, a corresponding working path is generated on the identified target working site; according to the working path, the mobile robot is instructed to work.

2. A mobile robot separate visual positioning and navigation method according to claim 1, characterized in that the step of identifying the target work site in the site image further includes calibrating the target work site and generating pixels. Mapping between points and actual points.

3. A mobile robot separate visual positioning and navigation method according to claim 2, characterized in that the preset working mode includes a patrol mode and a traversal mode.

4. A mobile robot separated visual positioning and navigation method according to claim 3, characterized in that when the preset working mode is the patrol mode, the corresponding work is generated on the identified target work site. Path steps include:

Determine the working point and working path according to the preset;

Convert the working point and working path into the corresponding set of pixel coordinates;

Set corresponding work action instructions at the work point according to the preset.

5. A mobile robot separated visual positioning and navigation method according to claim 3, characterized in that when the preset working mode is a traversal mode, the corresponding work is generated on the identified target work site. Path steps include:

Traverse and scan the site image to generate a zigzag working path;

Convert the zigzag working path into a corresponding set of pixel point coordinates.

6. A mobile robot separate visual positioning and navigation method according to claim 4 or 5, characterized in that, the step of instructing the robot to work according to the working path includes: reading the working path one by one, And instruct the mobile robot to operate according to the pixel coordinate set of the working path. At the working point, the corresponding work action instruction instructs the mobile robot to perform the corresponding work action;

Calculate the deviation value between the mobile robot and the current working path in real time. If the deviation value is greater than the preset maximum deviation value, instruct the mobile robot to return to the current working path.

7. A mobile robot separated visual positioning and navigation system, characterized by including a camera for real-time collection of corresponding site images, a control device, and a mobile robot for working according to instructions issued by the control device, so The camera is located above the target working site, and the control device is connected to the camera and the mobile robot signal;

The control device includes:

A target work site identification unit, used to identify the target work site in the site image, including identifying the boundaries and obstacles of the target work site;

The path generation unit is used to generate the corresponding working path on the identified target working site according to the preset working mode;

The control unit is used to instruct the mobile robot to work according to the working path.

8. A mobile robot separate visual positioning and navigation system according to claim 7, characterized in that the control device further includes a calibration unit for calibrating the target working site and generating pixel points corresponding to the actual Mapping between points.

9. A mobile robot separate visual positioning and navigation system according to claim 7 or 8, characterized in that the path generation unit includes:

The patrol path generation module is used to determine the working point and the working path according to the preset, convert the working point and the working path into the corresponding pixel point coordinate set, and then set the corresponding work action instructions at the working point according to the preset.

The traversal path generation module is used to traverse and scan the site image and generate a zigzag working path. diameter, convert the zigzag working path into a corresponding set of pixel point coordinates.

10. A mobile robot separate visual positioning and navigation system according to claim 9, characterized in that the control unit includes:

The instruction module is used to read the working path one by one, and instruct the mobile robot to operate according to the pixel point coordinate set of the working path. At the working point, the corresponding work action instruction instructs the mobile robot to perform the corresponding work action;

The judgment module is used to calculate the deviation value between the mobile robot and the current working path in real time. If the deviation value is greater than the preset maximum deviation value, instruct the mobile robot to return to the current working path.