JPS59205298A - Method of monitoring safety region of robot - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to monitoring the motion range of a robot arm, and particularly to a method of monitoring the motion range as an image.

[Background of the invention]

As a method for monitoring the operational safety range of a robot arm, for example, there is a method of attaching a plurality of ultrasonic transmitters/receivers to the arm to detect the approach of other arms and completely prevent collisions with the arms of other robots.

1) A method of installing fences etc. around the robot to prevent people from inadvertently entering the robot is also used. However, there is nothing that considers the safety of the operator during teaching. The ultrasonic method requires both ultrasonic transmitters and receivers, and the use of ultrasonic waves increases the cost of equipment and makes it impossible to determine the shape and color of obstacles. In addition, when considering safety measures for robots, there is a problem that if the robot cannot detect the shape of people or obstacles, it will not be a true safety measure.

Furthermore, in order to increase work efficiency, robot devices need to have faster arm movements. The robot's operation has become more complex with the diversification of usage methods, and the opportunities for sharing space with other equipment and human intrusion into the robot operating space are increasing. The entry of obstacles into the robot operating space is sudden and unpredictable from the viewpoint of robot control. From this point of view, the current situation is that no good countermeasure can be considered other than providing the robot controller with an obstacle detection function.

[Purpose of the invention]

The purpose is to detect the presence or absence of obstacles in the space in which the robot's arm is expected to move, as images, and to prevent the robot from colliding with obstacles.

[Embodiments of the invention]

Embodiments of the present invention will be described with reference to FIGS. 1 and 2. It is assumed that the robot arm 2 moves in all directions around the robot 1. Since the robot arm 2 makes arbitrary movements according to the scheduler 3, there is a risk of a collision occurring if an obstacle or a person enters the operating range 41 of the robot arm 2. The present invention is characterized in that when an obstacle enters the space within the robot's movement range, the robot is controlled so that it can perform safe movements.

The camera 9 constantly captures images of the operating range 41 of the robot arm 2 and stores them as surrounding environment data of the robot. Further, the camera 7 images the movement direction of the robot arm 2,
The space 40 is monitored to see if there are any obstacles. The details of the apparatus of the present invention will be explained below with reference to FIG.

Imaging of the motion range 41 of the robot arm 2 and its data processing will be described. The environmental data processing unit 15 controls the camera 9 via the camera orientation control unit 17 and the camera operation unit 8.
The area 41 is constantly scanned, and the input image from the camera 9 is written into the image memory 13 by the image input section 11. Each time image data is captured, the environmental data processing unit 15 compares the captured image data with the old forest land data corresponding to the captured image data, and if an obstacle is detected, it displays this on the console (or RTIO) and displays it to the operator. The operator determines whether the difference between the old environmental data and the new data is due to an obstruction, and inputs the information from the CRT.If it is determined that there is no obstruction, the environmental data processing mechanism updates the environmental data file. is replaced with new input data. Replacing the environmental data in this way is effective against slight environmental changes (for example, brightness that changes as time progresses).

The details of this process are shown in a flowchart in FIG. 3. The environmental conditions in the robot operating range without obstacles can be constantly updated and stored in the environmental data file 19, and can be used as basic data for detecting obstacles.

Next (・The mutual relationship between the arm 2 and camera 7 of this robot will be described. The robot arm 2 is controlled via the robot motion control mechanism 4 based on the information from the robot motion scheduler 3. On the other hand, the camera 7 is controlled by the robot motion control mechanism 4. It is controlled via the camera visual tracking section 5 and camera operation section 6 based on information from the scheduler 3, and is designed to always scan the front in the direction of movement of the arm 2. Image data is stored in the image memory by the image human power section 12. The image data stored in the image memory 14 is taken into the obstacle analysis section 16 and compared with the data in the environmental data file 19 to determine the presence or absence of obstacles in the space in the movement direction of the arm 2. The details of this process are shown in a flowchart in Fig. 4.The device of the present invention is structured in this way, so that it can immediately detect when an obstacle enters the space in the direction of movement of the robot arm 2, and the arm You can slow down or stop the movement of 2.

Further, when the surrounding brightness changes gradually, the environmental data is always replaced with the input from the camera 9, so that it does not affect obstacle detection.

Next, the structure of the fault space file will be described.

The obstacle space file subdivides the entire robot operation area into bases, and each base is divided into one bit of memory as shown in Figure 5(a).
Correspond as shown in (b). (a) shows the robot operation safe area, and (b) shows the obstacle space file. Set the focus point to '1', which corresponds to the point at which the failure occurs.This is how the robot moves, so before moving the robot, check whether there are any obstacles in the space it is trying to move through and execute the movement. This eliminates the risk of the robot arm colliding with obstacles.

Next, we will discuss the structure of environmental data. Environmental data is
Just like the obstacle space file, the entire robot movement area is subdivided into bases, and each base is stored in 8 bits of memory.
The correspondence is shown in Figure 6 (a) and (b). Therefore, the image data for each base can be stored as grayscale image data. Since the environmental data has such a configuration, by storing the environment in a safe state in all areas of robot operation in the form of a grayscale image, changes in the environment can be detected by comparing it with the grayscale image input by the camera. is possible. If this change is extremely small and is an overall change in brightness, it is determined to be a change in indoor brightness, and if there is a localized large change, it is determined that an obstacle has entered the area. do. Because of this, the device of the present invention allows the robot itself to operate while confirming its safety.

In the apparatus of the present invention, the following measures are taken into account when a robot obstacle enters an obstacle that is the same color as the background. Make it possible to change the color of the background part. For example, changing the back screen,
Maybe change the lighting only in the background. This change can be made periodically, and the environmental condition data is stored for each environmental condition of each bag.

This can be easily achieved, and
Any obstacle (it may be the same color as the background) can be detected. Note that the back conditions are changed periodically, and the period is determined by the speed of calculation processing. If the period is too large, it is not practical because it is impossible to respond quickly to the intrusion of obstacles. By making the main part of image processing hardware and speeding up the processing time, it is easily possible with current technology to reduce the processing time to about 0.1 to 0.2 seconds, so the back change time JjA is o, 2 to 1.0 seconds. It is appropriate to take about seconds. By performing K in this way, obstacles can be detected at a recognition speed comparable to that of human vision, allowing the robot to move safely.

There is a method of attaching ultrasonic waves to the robot arm to detect surrounding obstacles, but in this case, it is difficult to detect obstacles because the obstacles on the other side have complex shapes and the location is unknown. , many ultrasound devices must be attached to the arm. Therefore, the ultrasound equipment became an art university. In contrast, with the device of the present invention, the shape of the obstacle can be directly captured, so it is sufficient to install the minimum number of cameras required. For example, if it is sufficient to detect obstacles on a flat surface, it is sufficient to provide only two cameras. It is easy to understand that if you want to capture an obstacle three-dimensionally, you can do so by adding two more cameras, and the equipment cost is not that great. Additionally, by adding one more camera and zooming in to capture the obstacle, the location of the obstacle can be determined with higher precision.

This precision-improving camera has a suitable camera control mechanism, is designed to automatically capture obstacles in the scene, and repeatedly zooms up and amplifies the zoom until the desired precision is achieved.

Since the present device is structured in this manner, it is possible to recognize obstacles three-dimensionally and to detect the position of the obstacles with high precision. By doing so, it is also possible for the robot to perform precise operations on obstacles. Additionally, this device captures the obstacle (9) as a color gradation image (for example, 8 bits/pixel x 3 primary colors), so it can change the color tone at predetermined intervals even for obstacles with the same shape. It is possible to recognize them separately.

This allows you to colorize obstacles in a childish manner and have the robot process perform different actions accordingly. This device can recognize obstacles three-dimensionally, so by applying it to a multi-jointed robot, the multi-jointed arm can be effectively bent against dynamic obstacles and brought to the destination point. I can do it.

In order to achieve this purpose, ultrasonic systems require a huge number of ultrasonic transmitting and receiving devices on the arm, but this device can easily achieve this with a small number of camera devices. FIG. 7 shows an embodiment in which the device of the present invention is applied to an articulated robot. 1. is a robot, 2.
is an articulated arm, 42 is an obstacle, and 43 is an obstacle projection. Here, the obstacle is perceived as a flat surface, but as described above, by increasing the number of cameras, the object 42 can also be detected as a three-dimensional position. Thoughts (10
) is the same as the plane case, so it is omitted here.

The device of the present invention has the following advantages over conventional ultrasonic systems.

(1) Obstacles can be recognized three-dimensionally with simple equipment (with the ultrasonic method, it is necessary to add many ultrasonic transmitters and receivers to the arm, resulting in a beautiful equipment).

(2) Since the entire space can be divided according to the presence or absence of obstacles, the arm can be operated at high speed (with the ultrasonic method, it cannot be detected unless the obstacle is close to a certain extent,
It is necessary to limit the speed increase considering the inertia of the arm).

(3) Since the position and shape of obstacles can be recognized three-dimensionally, the arm movement path can be optimized (with the ultrasonic method, the only way to correct the movement path is after recognizing the obstacle) ).

(4) Since obstacles can be recognized not only by their shape but also by their color, it is also possible to know the specific nature of the obstacle (such as whether it is a human or another object).

(11) As described above, the device of the present invention has many advantages over the ultrasonic method. In this way, with the device of the present invention, using detailed information regarding obstacles, it is possible to control the robot completely safely depending on the purpose.

[Brief explanation of drawings]

Fig. 1 shows an example of the use and function of the present invention, and Fig. 2 shows an example of the use and function of the present invention.
The detailed structure of the device of the present invention is shown in FIGS. 3 and 4, the processing flow is shown in FIGS. An application example of the present invention is shown below. 1...Robot, 2...Robot arm, 3...
Robot motion scheduler, 4... Robot motion control unit, 5... Camera arm following unit, 6... Camera operation unit, 7... Camera (for imaging arm movement direction area), 8
...Camera operation unit, 9...Camera (for monitoring the entire robot operating area), 10...Console CRT, 11...
- Image input unit, 12... Image input unit, 13... Image memo 1c 14... Image memory, 15... Environmental data processing unit, 16... Obstacle analysis unit, 17... Camera orientation control unit, (12) 18...Obstacle space file, 19...Environmental data file, 20...Robot operation schedule 2, interface between robot operation control units, 21...Robot, robot operation control unit interface, 22...
Robot operation schedule 2, interface between camera arm following sections, 23... interface between camera arm following section, camera operation section, 24... interface between camera arm following section, image input section, 25... environmental data processing section , interface between camera control units, 2
6...Camera orientation control section, interface between camera operation sections, 27...Obstacle space file, interface between robot movement control sections, 28...Environmental data file, interface between obstacle analysis sections 5, 29... Failure analysis department, interface between failure space files, 30.
...Environmental data processing unit, interface between environmental data files, 31...Environmental data processing unit, interface between fault space files, 32...Environmental data processing unit, interface between image input units, 33...Image memory, Interface between obstacle analysis sections, 34...force (
13) camera, image input section, 35... camera operation section, interface between image input sections, 36... interface between image input section, image memory, 37... interface between image input section, image memory, 40. ...Robot arm movement direction area, 41...All robot arms (14) (b) Tree! -Month 4 Figure 6

Claims (1)

[Claims] 1. In a method of monitoring the movement area of a robot using images captured by a camera, at least a first camera images the movement area including the entire movement area of the robot arm, and another second camera monitors the movement area of the robot arm. imaging a partial operating region including the arm, changing the color tone of the imaging data by the first camera at a predetermined cycle;
Part 1. A method for monitoring a safe area of a robot, comprising detecting an obstacle based on a difference in image data obtained by a second camera.