Leveraging Deep Learning for Object and Position Recognition of Load Carriers for Autonomous Logistics Vehicles

This work presents a deep learning framework utilizing a convolutional neural network for the detection and pose estimation of load carriers in autonomous logistics vehicles, operating directly on RGBD images. The model identifies reference points on the carriers to compute their pose by integrating landmark data with geometric knowledge, demonstrating sufficient accuracy for industrial applications through extensive validation. This approach is significant for practitioners as it enhances the reliability of autonomous systems in intralogistics, facilitating efficient automated pickup operations.