CN105690371A - Space service robot-oriented hand-eye system - Google Patents

Abstract

The hand-eye system for space service robots has a main control pioneer robot platform (1), a robot self-positioning camera (2), and also includes a robotic arm (3), a mechanical claw (4) attached to the robotic arm (3) and A miniature camera (5), and an operation panel (6) of the simulated space capsule, a control button (7) on the operation panel (6), and a power supply module (8). The invention provides a vision-based mobile robot hand-eye system capable of assisting astronauts to complete certain tasks.

Description

A hand-eye system for space service robots

technical field

In order to meet the needs of using robots to assist astronauts to complete certain tasks in the simulated weightless environment in space, and to solve the problems of how service robots can identify targets, locate targets, and click buttons independently, the invention proposes and A vision-based visual feedback system for mobile robots is implemented. On the basis of image processing algorithms such as Hough transform and Cam-shift tracking algorithm, the features of the target image are extracted, and the visual measurement model established by the principle of pinhole imaging is used to obtain the mapping relationship between the image and the target and obtain the spatial information of the target point in real time , so as to control the manipulator to complete the specified task.

Background technique

With the continuous development of science and technology, more and more human beings will enter space. However, in the weightless conditions of space, the work of astronauts will be subject to many restrictions. As a result, astronauts can issue tasks and action commands through brain electricity, myoelectric biosignals, and voice signals, allowing space service robots to replace astronauts out of the cabin and complete designated tasks. In order to verify the above-mentioned multi-mode human-computer interaction interface technology, it is necessary to develop a service robot with certain intelligence. In a simulated space environment, the service robot autonomously operates the spacecraft control panel after receiving instructions from the astronauts. Therefore, how to make the service robot realize the task of autonomously identifying targets, autonomously locating targets, and autonomously clicking buttons is a problem that needs to be solved.

In terms of robots for space services, there are almost no patented inventions at home and abroad. The Mars rover Spirit and the Mars Opportunity launched from Cape Canaveral Air Force Station in the United States on June 10, 2003 are a kind of The space robot landed in Gusev Crater in the southern hemisphere of Mars on January 4, 2004. "Courage" is 1.6 meters long, 2.3 meters wide, 1.5 meters high and weighs 174 kilograms. Its "brain" is a computer capable of executing about 20 million instructions per second. Its main task is to detect whether there is water and life on Mars, and analyze its material composition to deduce whether Mars can be adapted for life through transformation. Both Spirit and Opportunity carried out reconnaissance and detection after landing, and did not serve astronauts in the space capsule. Therefore, our space-oriented service robot is very necessary for development. .

Contents of the invention

The object of the present invention is to design a mobile robot hand-eye system based on monocular vision by utilizing the Pioneer mobile robot system. Through digital image processing methods such as Hough transform and Cam-shift tracking algorithm, the target image features are extracted, combined with prior knowledge, and the mathematical model established by the principle of pinhole imaging is used to obtain the spatial information of the target in real time, so as to complete the target detection. The task of identifying and guiding the manipulator to click a button. Describe in detail the technical content of the invention:

Referring to Figure 1, the robot autonomously recognizes the target according to the given task, autonomously locates the target, and autonomously clicks the button. Figure 1 is a block diagram of the position control principle of the vision system. The system extracts the coordinate information of feature points on the two-dimensional image plane from the image through image acquisition and feature extraction. Then use the geometric model of the target and the camera to estimate the positional relationship between the target and the camera. Then, the error between the estimated value and the ideal position is transmitted to the robot controller, and the controller uses the error to calculate the control signal of the manipulator, so as to control the manipulator to complete the task.

In this paper, based on the depth information obtained by using the target image obtained by the camera under the condition of known target information, a monocular vision distance measurement model is established, and then the physical shape and color features of the target are extracted. Because it is through the recognition of color features and contours, the accuracy of recognition is improved. Before feature extraction, the target image is smoothed, dilated and corroded, combined with feature attributes. Dilation is convolving an image with a kernel, typically a small solid square or disk with a reference point in the middle. Erosion is the inverse operation of dilation, which computes the minimum value of pixels in the kernel region. First, use the image color space conversion function cvCvtColor to convert the collected RGB image into a grayscale image. In order to eliminate the interference of high-frequency noise on the detection circle, the filter function cvSmooth is used to perform Gaussian convolution with a kernel size of 9×9 on the image. Finally, use the Hough circle transform cvHoughCircles function to detect circles on the image. The cumulative resolution of finding the center of the arc is set to 6, the minimum distance between two different circles is set to 120, the upper threshold for the edge is set to 36, the threshold of the accumulator is set to 85, and the minimum circle radius is set to 3 , set the maximum circle radius to 250.

Compared with the prior art, the advantages and effects of the present invention should preferably be analyzed from the structure and have appropriate data.

The technical problem to be solved by the patent of the present invention is to provide a vision-based mobile robot hand-eye system that can assist astronauts to complete certain tasks.

This system has a main control pioneer robot platform, and also includes a robotic arm, a miniature camera attached to the robotic arm, an analog space capsule operating panel, control buttons on the operating panel, and a power module. The robot platform adopts the P3-DX Pioneer robot, which runs the windowXP system internally as the control core, and the peripheral configuration is connected with the corresponding mechanical structure and electrical structure to realize the overall hand-eye system. Robai's seven-free manipulator uses the CytonC++ API provided by it to build a manipulator control platform. The miniature camera adopts a Canon high-definition camera with a diameter of less than 1 cm, which is connected to the robot through a USB port to realize image acquisition and transmission of the operation panel. The space capsule operation panel and the control buttons on the panel are developed and laid out by ourselves. The 6 buttons are 6 colors, which are used for identification and control inside the system. The entire operation panel is used to simulate the environment in the space capsule.

Attached drawings and their brief descriptions:

Figure 1 is a block diagram of the position control principle of the visual system of the system;

Figure 2 is the overall appearance of the system;

Fig. 3 is the space capsule operation board of this system;

Example:

The hand-eye system for the space service robot of the present invention realizes the tasks of autonomously identifying targets, autonomously locating targets, and autonomously clicking buttons according to a given task.

Referring to Figure 1, the system extracts the coordinate information of feature points on the two-dimensional image plane from the image through image acquisition and feature extraction. Then use the geometric model of the target and the camera to estimate the positional relationship between the target and the camera. Then, the error between the estimated value and the ideal position is transmitted to the robot controller, and the controller uses the error to calculate the control signal of the manipulator, so as to control the manipulator to complete the task.

Referring to the robot shown in Figure 2, in order to meet the task requirements, on the main hardware of the system, a seven-degree-of-freedom manipulator is mounted on the Pioneer mobile robot system, and a hand-eye system is constructed by installing a small camera at the end of the manipulator.

Referring to Figure 3, there are 6 buttons and a knob on the operating panel of the space capsule. After the robot is given the command to click buttons of different colors, the detection target is identified by double feature extraction of shape and color, and the recognition result is calibrated. .

Claims (3)

1. The hand-eye system for space service robots, with the main control pioneer robot platform (1), the robot self-positioning camera (2), and also includes the mechanical arm (3), the mechanical claw (4) attached to the mechanical arm (3) ) and a miniature camera (5), as well as the control button (7) on the simulated space capsule operating panel (6) and the operating panel (6), and the power supply module (8). Vision-based mobile robot hand-eye system. 2. The hand-eye system for space service robots according to claim 1, characterized in that: the mechanical arm (3) adopts 7 degrees of freedom, is loaded directly above the main control robot platform (1), and can comes into contact with the front of the gripper (4). 3. The hand-eye system for space service robots according to claim 1, characterized in that: the miniature camera (3) is installed directly above the end of the mechanical arm (2), so that the robot can photograph the mechanical claw (4 ) of the front image and identify it.