Connecting Navigation to Manipulation

A mobile robot can reach the right location and still be poorly positioned to grasp an object. For a ROS 2 course project, I implemented a custom A* global planner and a trajectory-feedback controller as Nav2 plugins, then helped connect navigation to the arm’s approach sequence.

The planner searched the occupancy grid; the local controller corrected cross-track and heading errors. A finite-state machine coordinated navigation, target approach, grasping, and return.

Making the Final Approach Reachable

During the approach, ArUco observations replaced accumulated odometry as the local target reference. Repeated inverse-kinematics queries determined when the object entered the arm’s reachable workspace before triggering the grasp.

The integrated system demonstrated navigation, object localization, and grasping, including trials with randomly placed targets. Implementing the planning and control modules inside Nav2 kept them usable within the full robot stack.

Robot navigation and arm control demonstration