Real-Time Face-Tracking Control of a 4 DOF Manipulator Using Deep Learning Detection

This paper has presented the design, simulation, and real-time implementation of a four-degree-of-freedom (4-DOF) robotic system for face detection and tracking. The system has combined kinematic control with deep learning. It has used the Xception network for high-accuracy facial detection. A Fuzzy PID-MIJ control algorithm has optimized the trajectory planning. This has ensured smooth and precise target tracking. Both simulations and physical experiments have validated the system's real-time performance. The results have shown that the system can reliably track faces with minimal error. This approach has demonstrated significant potential for dynamic, human-interactive applications.