Project Overview
Robotic arm application that collaboratively operates with human workers in a shared workspace. The application involves Motion Planning and Object & Human parts (torso, arms) Recognition.
The end product has been ported with a safety layer that stops the robot at the instant there is an unpredictable force given to the joint ESCs. The perception system recognizes a human worker around it with perfect accuracy (with a non-learning image recognition method, this is ensured, but with more assumptions in the workspace).
After human worker leaves the workspace, the robot continues its groovy Pick & Place task, if the condition for picking still satisfies. Otherwise it goes back to its rest pose.
Hardware & Technologies
- UR5 Manipulator Arm: Primary robotic hardware for manipulation tasks.
- Kinect 2: Depth sensing for workspace monitoring and human detection.
- ROS Kinetic: Middleware for system integration and communication.
- OpenCV: Computer vision library for object and human recognition.
- Gazebo: Simulation environment for testing safety protocols.