Autonomous vehicles must detect nearby road users quickly and reliably, even when objects are partly hidden. We developed a lightweight camera-based method that adapts its processing to scene difficulty, combines information across video frames, and learns from laser-sensor data during training. It reached 38.7% on the benchmark detection measure and processed 26.2 frames per second without laser sensors during use, showing strong potential for fast, lower-cost vehicle perception.

Aiming at stick failures under hard‑soil excavation, this paper presents a mechanism‑structure collaborative optimization for a 20‑ton excavator. With test‑identified hazardous conditions and Kriging surrogate models, NSGA‑II‑based two‑level optimization boosts digging force by 9 %, cuts peak stress by 7 % and reduces stick weight by 9 %. It provides references for excavator‑attachment multidisciplinary optimization.

Robotic grasping in complex environments is challenging due to variations in object positions, orientations, and backgrounds. This study proposes a vision-based method for collaborative robots that integrates object detection, target mask generation, and grasp position estimation to improve grasping reliability. Experiments on a robot platform demonstrate effective target-specific grasping in unstructured multi-object scenes, showing its potential for intelligent manufacturing applications.