Autonomous Maze Navigation with 6DOF Robotic Arm

This project demonstrates a fully integrated workflow that bridges digital simulation and real-world robotics to achieve reliable maze navigation. By leveraging vision-based calibration with ArUco markers to translate image pixels into physical coordinates, implementing a BFS path-planning algorithm to determine optimal waypoints, and using MATLAB’s inverse kinematics via a Python interface to generate precise joint trajectories, the MyCobot Pro 600 reliably traverses a 4×4 maze both virtually and in reality. The close alignment between the digital twin’s simulated motions and the robot’s actual performance validates the accuracy of the calibration and control pipeline. Overall, this work underscores the effectiveness of combining computer vision, algorithmic planning, and kinematic modeling for autonomous robotic tasks, and provides a scalable foundation for tackling more complex navigation challenges in future applications. 

Here the simulation and actual working hardware is mentiond in the images.

The images of MyCobot 600 performing the task is mentiond here.

Full working model with hardware implementation in there is the video.  

You can find more detailed report for this project here.👉 Link