Controlling a 4DOF Robotic Arm

For this series of projects in my WPI RBE 500 and 501 courses, my teams and I created different controllers to actuate a 4DOF robotic arm with a gripper.

In this project, we used ROS2 Python and MATLAB (in separate portions) to command the robot. I took the lead in deriving inverse kinematics of the robot from CAD models and Jacobian matrices for accurate robot motion planning. The robot was successfully programmed to grab and move an object, move between points while avoid an obstacle space, and translate the end-effector in a single linear axis despite not having linear actuators.

Trajectory planning was also implemented, and my team and I created quintic-polynomial-based and Linear Segmented Parabolic Blend (LSPB) trajectories for the robot to move between waypoints. Additionally, LSPB trajectories were "stitched together" in a manner such that the robot maintained the trapezoidal velocity profiles while not reaching zero velocity in all joints at the intermediary waypoint in obstacle avoidance. Finally, controllable dynamics of the robot were calculated using the mass moment of inertia matrices for each of the robot links and motors, and the robot was used to resist and estimate a wrench applied to the end effector.