The robot DARWin developed with NI LabVIEW was the first in the United States to enter the "Robot World Cup Football Tournament".
Author:
Karl Muecke-Virginia Tech University
industry:
University/Education, Research, Mechatronics/Electrical Engineering
product:
LabVIEW, machine vision, real-time module
challenge:
Develop a robot system as a new gait research platform, and design the first humanoid robot in the United States to participate in RoboCup (autonomous robot soccer game).
solution:
The solution uses NI LabVIEW to interact with third-party hardware to accelerate the development and testing of new gait research; at the same time, it uses LabVIEW real-time module and NI vision software to create artificial intelligence, so that the robot can perform advanced functions, such as playing football.
"Using LabVIEW's gamepad VI can easily realize this idea. We can easily debug the gait and movement without having to debug the robot movements at the same time."
To study bipedal gait, Virginia Tech Robotics and Machinery Laboratory (RoMela) designed the original intelligent dynamic anthropomorphic robot (DARwIn). At RoMela, we tested the assumptions and theories of gait research on hardware. We decided to use RoboCup, an international robot football game, as a stage to demonstrate the superiority and feasibility of this robot, and to demonstrate the robustness of DARwIn.
The current robot programming and control technology usually uses C code, which has a steep learning curve and relatively difficult hardware update. At RoMeLa, we use NI technology to accelerate the development of new robot gait, and finally make a fully autonomous humanoid robot come out. The robot can not only play football, but also serve as a research platform for new gait. We use the LabVIEW graphical development platform not only to create scalable and adaptable software, but also to develop robot brains that perform advanced tasks, such as playing a football game.
LabVIEW as an extensible hardware interfaceDue to the variety of robot platforms in RoMeLa, we need to choose a system that can adapt to different hardware settings through simple configuration. Most small robotics research uses personal digital assistants (PDAs) to autonomously control robots. After adopting the LabVIEW real-time module on the PC104+ computer, the computer architecture can be expanded almost without any overhead. Unlike a PDA, the system is compatible with a range of different sensors-IEEE 1384 cameras, RS485 communications, multiple wireless networks and many other devices. Adding a new camera or 802.11 port, adapting driver, writing C or C++ code takes a lot of time, and LabVIEW VI makes all the problems easy to solve.
At present, we use LabVIEW to control the robot movement through RS485, and read the joint position from the built-in potentiometer of the servo motor on the same serial network. While the robot is walking or moving, the gyroscope with acceleration and direction information rate communicates with LabVIEW through the RS232 serial port. At this time, the program can modify the walking gait in real time to efficiently control the robot balance.
The initial robot development platform only required servo motors and rate gyroscope interfaces. However, the hardware and programming required to participate in the RoboCup competition in July 2007 is undoubtedly more complicated. In addition to the robot walking and maintaining balance, the software must also provide vision, brain, and communication functions. Since in RoboCup competitions, robots must be fully autonomous and controlled wirelessly, only the network host can be used to control the start/stop signals of the robot. LabVIEW runs all robot software in real time, reducing cost and CPU time.
Validated gait researchWhen generating mathematical formulas for robot gait, it is difficult to visualize the results. Using LabVIEW can not only call the gait generated by other calculation software packages (such as Wolfram Mathematica or Microsoft Excel), but also help to study the gait by establishing the visual motion graphics of the robot. Using LabVIEW's 3D graphics control, we can simulate the behavior of the robot when it executes the generated gait. Since there is no need to use physical hardware to set up and test the gait, a lot of research and development time is saved.
When testing the gait of the robot on physical hardware, we can imitate the artificial intelligence of the robot through the gamepad controlled by the user. The user acts as the eyes and brain of the robot, and the handle acts as an interface to control the robot's actions, sending commands to control the robot's walking, kicking, and fighting actions. Using LabVIEW gamepad VI, this idea can be easily realized. We can easily adjust the gait and movement without the need to adjust the robot movements simultaneously.
Create artificial intelligenceWithout any pre-visual processing, a graduate student can configure two IEEE 1394 cameras and write a VI program (for the relative position recognition and physical positioning of the orange football used by RoboCup) in just two hours. To accomplish this task, any other college in the world may require many students to spend a long time studying the code. But with our development efficiency, a student can achieve DARwIn football action control in just one week. This robot is the first and only humanoid robot in the United States to participate in the RoboCup competition.
Dongguan Tuojun Electronic Technology Co., Ltd , https://www.fibercablessupplier.com