The field of construction and mining is witnessing significant advancements in autonomous systems, driven by the integration of computer vision, robotics, and deep learning. Researchers are exploring innovative approaches to enhance productivity, reduce operational costs, and improve safety in aggregate handling, rock particulate classification, and regolith excavation. A key direction in this field is the development of adaptive and compliant control systems that can effectively interact with complex environments and diverse materials. The use of reinforcement learning and attention mechanisms is also gaining traction, enabling autonomous systems to learn from experience and improve their performance over time. Noteworthy papers in this area include:

• A study on autonomous aggregate sorting using a computer vision-aided robotic arm system, which achieved an average grasping and sorting success rate of 97.5%.
• A deep learning model for rock particulate classification, which introduced self-attention and channel attention mechanisms to improve classification accuracy and robustness.
• A framework for learning tool-aware adaptive compliant control for autonomous regolith excavation, which demonstrated the importance of procedural generation and visual feedback in developing robust and versatile autonomous systems.
• An RL framework for interactive shaping of granular media, which enabled a robotic arm to shape granular media into desired target structures with high accuracy.