Project Description

This project was performed for my ME 532 Robot Modeling class

The project involves working with a team to model, simulate, and control a 6-DOF UR5 robotic arm, able to perform autonomous pick and place tasks in a virtual environment, incorporated with obstacle avoidance and user defined functions. 

The robots was created by importing the UR5 kinematic model into MATLAB using the robot modeling add-on. First verifying it had proper collision geometry for safe motion planning. Next a simulated environment was created in a 3D environment, complete with a table surface, a center wall, and two objects to interact with. A top-down grabber was added and using inverse kinematics to create a grasping routine with a simple two finger gripper. The end effector is then able to approach the object and "grab" it by computing a relative transform and carrying it by updating the transform as it moves. Path planning and obstacle avoidance where added and incorporated into a user interface, where the system user is able to input custom target coordinates. This then allows the code to check the coordinates, plan a collision free trajectory, and return an error if the object is un-reachable. 

Incremental waypoint-based inverse kinematics where used, which solve each segment of the path using the previous solution as the initial guess. This then avoids jerky movements that comes from solving inverse kinematics in one large jump. The addition of verified trajectories which were performed before executing any motion, ensures the arm never contracts the wall or itself. 

Overall this project deepened my understanding of inverse kinematics, motion planning, and robot simulation. It gave me a first had expereince in coding robots, applying the academic concepts to a real-life simulation, verifying the robots motion in real time.