This study presents an autonomous navigation system for grape cultivation robots in overhead trellis vineyards, combining two navigation methods: YOLOv8-based detection of QR codes which uniquely identify grape clusters and provide cultivation data for precise positioning, and Global Navigation Satellite Systems (GNSS) technology for broader vineyard navigation. Testing in a simulated environment demonstrated 83.3% accuracy in object detection-based positioning and 90% success in GNSS navigation within 1.0 m of targets. The system's dual-approach navigation shows promise for addressing agricultural labor shortages while improving grape cultivation efficiency.