Aiming at the problem that the existing neural network models have insufficient ability to characterize the features of classification objects in image classification tasks and cannot achieve high recognition accuracy, an image classification algorithm based on Lightweight Group-wise Attention Module (LGAM) was proposed. The proposed module reconstructed the feature maps from the channel and space of the input feature maps. Firstly, the input feature maps were grouped along the channel direction, and channel attention weight corresponding to each group was generated. At the same time, ladder type structure was used to solve the problem that the information between the groups was not circulated. Secondly, the global spatial attention weight was generated based on the new feature maps concatenated by each group, and the reconstructed feature maps were obtained by weighting the two attention weights. Finally, the reconstructed feature maps were merged with the input feature maps to generate the enhanced feature maps. Experiments were performed on the Cifar10 and Cifar100 datasets and part of the ImageNet2012 dataset with using the classification Top-1 error rate as the evaluation indicator to compare the ResNet, Wide-ResNet and ResNeXt enhanced by LGAM. Experimental results show that the Top-1 error rates of the neural network models enhanced by LGAM are 1 to 2 percentage points lower than those of the models before enhancing. LGAM can improve the feature characterization ability of existing neural network models, thus improving the recognition accuracy of image classification.

In order to study the intelligent voice control of aircraft towing vehicle, realize accurate and efficient recognition of the voice command of pilot in the airport environment, and solve the problems of large computation, high time complexity and low recognition efficiency of the traditional Dynamic Time Warping (DTW) algorithm, a new optimization algorithm of DTW with constraint of hexagonal warping window for vehicle speech recognition was proposed. Firstly, the influence of warping window on the accuracy and efficiency of DTW algorithm was analyzed from three aspects such as the principles of DTW algorithm, the speech characteristics of towing vehicle instruction and the airport environment. Then, on the basis of DTW optimization algorithm with constraint of Itakura Parallelogram rhombic warping window, a DTW global optimization algorithm with the constraint of hexagonal warping window was further proposed. Finally, by varying the optimization coefficient, the optimal DTW algorithm with the constraint of hexagonal warping window was realized. The experimental results based on isolated-word recognition show that, compared with the traditional DTW algorithm and the DTW algorithm with rhombic warping window constraint, the recognition error rate of the proposed optimal algorithm is reduced by 77.14% and 69.27% respectively, and its recognition efficiency is increased by 48.92% and 27.90% respectively. The proposed optimal algorithm is more robust and timeliness, and can be used as an ideal instruction input port for intelligent control of aircraft towing vehicle.