【摘要】 针对印刷仓储搬运过程中液体油墨等有害物质倾斜泄露导致人身安全及环境安全等诸多问题，本研究提出了一种基于陀螺仪技术的姿态调整算法（FRPC），即基于比例-积分-微分的快速响应预测控制算法。首先，提出一种基于并发陀螺仪的注意力网络（CGAN），用以处理多数据并发流，通过并行策略和动态关注度调整，自动分配计算资源，提高数据处理效率；其次，根据陀螺仪数据存在时序性特点，设计了一种基于时间序列的注意力并发网络（TSACN），该算法结合时间序列专用神经网络结构，以当前及历史数据预测姿态变化，进一步提高了调整的准确性与效率；最后，在姿态调整算法和姿态稳定架构的相互配合下，实现了物品姿态的精准监测与快速校正。在实验环境下，本研究所提出的基于陀螺仪技术的姿态调整算法架构设计可快速精准地校正姿态，保证印刷仓储搬运过程中物品运输安全。更多还原

【Abstract】 Due to liquid ink and other hazardous substances tilt leakage leading to personal safety and environmental safety, and many other problems in the printing storage and handling process, a gyroscope-based attitude adjustment algorithm(FRPC) was proposed. Firstly, Concurrent Gyroscopic Attention Network(CGAN) was proposed to deal with the concurrent flow of multiple data, and automatically allocated computational resources to improve the data processing efficiency through the parallel strategy and dynamic attention adjustment. Secondly, based on the temporal characteristics of gyroscope data, Time Series Attention Concurrent Networks(TSACN) was designed, which combined the time series-specific neural network structure with current and historical data to predict attitude changes, further enhancing the accuracy and efficiency of adjustment. Finally, with the mutual cooperation of FRPC and attitude stabilisation architecture, the accurate monitoring and fast correction of object attitude were achieved. In the experimental environment, the gyroscope-based attitude adjustment algorithm architecture design proposed in this study can quickly and accurately correct the attitude, so that the safety of the printing and storage handling process can be guaranteed.更多还原