In 3D reconstruction tasks, single-sensor data fusion based on deep learning is limited by the integrity and accuracy of the data, which reduces the accuracy and reliability of the fusion results. To address this issue, this study proposes a multi-sensor data fusion method based on a self-attention mechanism. A multi-sensor data fusion model for acquiring multi-source and multi-modal data is constructed, with the core component being a convolutional neural network with self-attention (CNN-SA), which employs CNNs to process multi-source and multi-modal data by extracting their features. Additionally, it introduces an SA mechanism to weigh and sum the features of different modalities, adaptively focusing on the importance of different modal data. This enables mutual support, complementarity, and correction among the multi-modal data. Experimental results demonstrate that the accuracy of the CNN-SA network is improved by 72.6%, surpassing the improvements of 29.9% for CNN-CBAM, 23.6% for CNN, and 11.4% for CNN-LSTM, exhibiting enhanced generalization capability, accuracy, and robustness. The proposed approach will contribute to the effectiveness of multi-sensor data fusion processing.

Owing to the assembly state changes during aircraft assembly processes, assembly force-deformation problem occurs. To obtain the structure shape in the product assembly process efficiently and accurately, a three-dimensional (3D) mapping technology for the structural deformation during the aircraft assembly process is proposed combined with a fiber Bragg grating (FBG) optical fiber sensor and binocular vision measurement system. First, this study established a curvature transformation model using optical fiber monitoring data, obtained the 3D spatial deformation of the product, and completed the unification of the optical fiber wavelength change and spatial 3D point coordinate heterogeneous data. Second, a mesh deformation optimization algorithm based on point-cloud optimization was established. Subsequently, the deformation effects of four mesh deformation models were compared to verify the feasibility and accuracy of HEC-Laplace, and the 3D mapping of the product structure shape in the assembly process was realized. Finally, a cantilever wing model was used to verify the deformation of different loading modes. The results show that the product structure changes can be accurately obtained through the proposed technology, thereby improving the accuracy control and overall assembly quality in the aircraft assembly process and providing a theoretical basis for intelligent aircraft assembly.

This study proposes a multi-objective sensor layout optimization method combining weight factor and sensing coverage, which aims to solve the problem that the uneven distribution of sensor layout for deformation monitoring of large composite structures in assembly leads to redundancy of monitoring coverage and fails to accurately reflect the global deformation of structures. First, a modal metric function model was established to determine the optimal number of sensors, and the optimal modal function was identified for layout optimization. Second, a functional model incorporating perceptual coverage, weight factor, and overlap was established. Using the weight factor and perceptual overlap as constraints and aiming for optimal modal and perceptual coverage rate, a statistical method was employed to allocate deformation weight factors, and the perceptual range was introduced to optimize layout angles, addressing uneven sensor distribution and redundant coverage while optimizing sensor placement. Finally, the feasibility of structural deformation monitoring was validated through comparative analysis by integrating the vision measurement system based on the optimization results. The results demonstrated that the relative error of deformation monitoring was within 6.47%, the root mean square error was 0.066 mm, and the reconstruction error rate was reduced by 26.16%. Based on the premise of lightweight sensors, this method allocates weight factors according to deformation trends, optimizes sensor positions and angles in combination with sensing coverage, effectively ensures the reliability of global deformation monitoring, and offers a novel approach for optimizing damage evolution sensor layouts.

Thermal protection systems (TPSs) are important components of reusable spacecraft, and their assembly quality has a crucial impact on flight safety. Owing to the complex assembly process and variable states of spacecraft thermal protection systems, assembly parameters may vary under different assembly states. Therefore, to obtain assembly parameters accurately and efficiently under different assembly states, in this study, 3D point cloud data and fiber optic sensor data were fused to develop an assembly parameter update method for assembly process state changes. Firstly, based on the measured data of thermal protection components and load-bearing structure, the gap, flush and matching parameters solution model are proposed. Secondly, to address the deformation problem of the load-bearing structure caused by changes in assembly status, a fusion method based on laser scanning and sensor detection was devised to achieve deformation prediction of the assembly structure during the assembly process. Thirdly, based on the assembly parameter solution model and point cloud prediction model, a constraint-based assembly parameter optimisation model was established, and an improved quantum particle swarm optimisation (LQPSO) algorithm was employed to achieve assembly parameter updates oriented toward changes in assembly status. Finally, an experimental system for array-based thermal protection structure simulation was established to validate the proposed method. The results show that the proposed parameter update method can achieve ideal results for different assembly state simulation components.

In the large-scale measurement field, deployment planning usually uses the Monte Carlo method for simulation analysis, which has high algorithm complexity. At the same time, traditional station planning is inefficient and unable to calculate overall accessibility due to the occlusion of tooling. To solve this problem, in this study, we first introduced a Poisson-like randomness strategy and an enhanced randomness strategy to improve the remora optimization algorithm (ROA), i.e., the PROA. Simultaneously, its convergence speed and robustness were verified in different dimensions using the CEC benchmark function. The convergence speed of 67.5–74% of the results is better than the ROA, and the robustness results of 66.67–75% are better than those of the ROA. Second, a deployment model was established for the large-scale measurement field to obtain the maximum visible area of the target to be measured. Finally, the PROA was used as the optimizer to solve optimal deployment planning; the performance of the PROA was verified by simulation analysis. In the case of six stations, the maximum visible area of the PROA reaches 83.02%, which is 18.07% higher than that of the ROA. Compared with the traditional method, this model shortens the deployment time and calculates the overall accessibility, which is of practical significance for improving assembly efficiency in large-size measurement field environments.

In complex measurement systems, scanning the shape data of solid models is time consuming, and real-time solutions are required. Therefore, we developed a 3D non-rigid transformation deduction model based on multi-control point perception data. We combined a convolutional neural network (CNN), gated recurrent unit (GRU), and self-attention mechanism (SA) to develop the CNN-GRU-SA deduction network, which can deduce 3D non-rigid transformations based on multiple control points. We compared the proposed network to several other networks, with the experimental results indicating that the maximum improvements in terms of loss and root-mean-squared error (RMSE) on the training set were 39% and 49%, respectively; the corresponding values for the testing set were 48% and 29%. Moreover, the average deviation of the inference results and average inference time were 0.55 mm and 0.021 s, respectively. Hence, the proposed deep learning method provides an effective method to simulate and deduce the 3D non-rigid transformation processes of entities in the measurement system space, thus highlighting its practical significance in optimizing entity deformation.