Cooperative Assembly in 3C Manufacturing with Micron-Level Precision Using Distributed Neural Networks

Manual labors are commonly adopted in the assembly of 3C components due to their complex structure, small pitch, and susceptibility to damage. These characteristics have led to high worker strain and inconsistent quality in manual assembly. While high-precision industrial robotic arms can meet assembly requirements, the high costs of robots have limited their practical application. This paper proposes a collaborative macro-micro robotic system for 3C assembly tasks, where a low-cost 3-PRR parallel mechanism is utilized to position the 3C components with high precision. The positioning process of the connector is divided into two stages by the developed assembly strategy, where a Siamese neural network and an improved Yo1ov8 model are incorporated to recognize and align the assembly. This strategy is then applied to complete the assembly with a specifically designed macro-micro robot. Moreover, an SSA-LSTM model is developed to predict the force during the pressing and fitting of the connectors, thereby evaluating the assembly quality. A distributed neural network model is adopted across various stages of the assembly, significantly enhancing the reliability and efficiency of this task. The high-precision strategies and low-cost structure proposed in this paper can provide theoretical and technical contributions to achieving efficient“ precise” and flexible assembly goals.