In order to address the issue of time-consuming, labor-intensive traditional industrial tomato sorting, this paper proposes a high-precision tomato recognition strategy and fast automatic grasping system. Firstly, the Swin Transformer module is integrated into YOLOv10 to reduce the resolution of each layer by half and double the number of channels, improving recognition accuracy. Then, the Simple Attention Module (SimAM) and the Efficient Multi-Scale Attention (EMA) attention mechanisms are added to achieve complete integration of features, and the Bi-level Routing Attention (BiFormer) is introduced for dynamic sparse attention and resource allocation. Finally, a lightweight detection head is added to YOLOv10 to improve the accuracy of tiny target detection. To complement the recognition system, a single-vertex and multi-crease (SVMC) origami soft gripper is employed for rapid adaptive grasping of identified objects through bistable deformation. This innovative system enables quick and accurate tomato grasping post-identification, showcasing significant potential for application in fruit and vegetable sorting operations.

This paper proposes a soft grasping system with stiffness self-perception and adaptive adjustment capabilities. The sensing unit consists of liquid optical lenses. It uses liquid pressure to drive film deformation to form an equivalent convex lens. It determines the width, mass, surface stiffness, and other properties of the target based on the voltage fluctuations caused by focal length variations. The variable stiffness layer operates on the air-quicksand variable stiffness principle. It achieves a wide range of stiffness adjustment through the regulation of positive and negative air pressure. Based on tactile sensing information, an adaptive grasping strategy matching the physical properties of the target with the stiffness of the gripper is established, with the surface stiffness and external dimensions of the target as the classification basis. Target property identification and classification are completed through the voltage signals of the sensor. The set value of the gripper stiffness is determined according to the action priority including compliant envelope, stable grasping, and anti-interference performance. The precise control of stiffness and driving pressure is further realized through air pressure regulation. Sorting and occluded target sorting experiments are designed to verify that the proposed strategy can stably complete multi-target recognition and grasping tasks without relying on visual feedback. This work provides a new approach to solving the perception and adaptation challenges of soft grasping systems in complex environments.

This paper focuses on improving the tracking accuracy for servo systems and increasing the contouring performance of precision machining. The dynamic friction during precision machining is analyzed using the LuGre model. The dynamic and static parameters in the friction model are efficiently and accurately identified using the improved Drosophila swarm algorithm based on cross-mutation. The friction tracking error can be deduced from the friction state space and an expression is derived. To compensate for the tracking error caused by friction, a feedforward compensation control is designed to avoid signal lag in traditional friction controllers. Furthermore, the factors of multi-axis parameter mismatching that impact the machining profile accuracy are analyzed for multi-axis control. An adaptive cross-coupled control-based pre-compensation strategy of contour error is designed to reduce both the tracking error and the contour error. The effectiveness of the proposed method is validated through several experiments, which demonstrate a remarkable improvement in tracking performance and contour accuracy.