Integrating FingerVision to Robotic Systems

Here is an example of integrating FingerVision to a Baxter robotic system. This Baxter has a Baxter electric parallel grippers (EPG) and a Robotiq 2-Finger Adaptive Robot Gripper. Each gripper has two fingers, and on each finger a FingerVision is installed. Thus the robot has four FingerVision sensors.

For the integration, we are using two Raspberry Pi 3 Model B. RPi captures images, and streams them in MJPG format through Gigabit Ethernet network. MJPG-streamer is useful for streaming on Raspberry Pi.

Note: We tried https://github.com/jacksonliam/mjpg-streamer and https://svn.code.sf.net/p/mjpg-streamer/code/mjpg-streamer , and found that the GitHub version is better in our setup.

Note: We are using Ubuntu MATE 16 (ubuntu-mate-16.04-desktop-armhf-raspberry-pi.img.xz) as the operating system of RPi. We found that a newer version (Ubuntu MATE 16.04.2) had a trouble in streaming MJPG data; it seemed that ffmpeg installed in default was the reason of this issue. Use Ubuntu MATE 16.04.

A central PC is used to process the images, including marker tracking and proximity vision. The ROS package (see Software#ROS) is used. We could process data at more than 30 fps with 320x240 resolution for each FingerVision.