Aiming at the problem that it is difficult to distinguish similar land types in outdoor scenes with multiple objects and complex spatial topological relationships， an A-Edge-SPG （Attention-EdgeConv SuperPoint Graph） graph neural network combining graph model and attention mechanism module was proposed. Firstly， the superpoints were segmented by the combination of graph cut and geometric features. Secondly， the local adjacency graph was constructed inside the superpoint to capture the context information of the point cloud in the scene and use the attention mechanism module to highlight the key information. Finally， a SuperPoint Graph （SPG） model was constructed， and the features of hyperpoints and hyperedges were aggregated by Gated Recurrent Unit （GRU） to realize accurate segmentation among different land types of point cloud. On Semantic3D dataset， the semantic segmentation effect of A-Edge-SPG model and SPG-Net （SPG neural Network） model was compared and analyzed. Experimental results show that compared with the SPG model， A-Edge-SPG model improves the Overall segmentation Accuracy（OA）， mean Intersection over Union （mIoU） and mean Average Accuracy （mAA） by 1.8， 5.1 and 2.8 percentage points respectively， and significantly improves the segmentation accuracy of similar land types such as high vegetation and dwarf vegetation， improving the effect of distinguishing similar land types.

In the framework of traditional knowledge distillation， the teacher network transfers all of its own knowledge to the student network， and there are almost no researches on the transfer of partial knowledge or specific knowledge. Considering that the industrial field has the characteristics of single scene and small number of classifications， the evaluation of recognition performance of neural network models in specific categories need to be focused on. Based on the attention feature transfer distillation algorithm， three specific knowledge learning algorithms were proposed to improve the classification performance of student networks in specific categories. Firstly， the training dataset was filtered for specific classes to exclude other non-specific classes of training data. On this basis， other non-specific classes were treated as background and the background knowledge was suppressed in the distillation process， so as to further reduce the impact of other irrelevant knowledge on specific classes of knowledge. Finally， the network structure was changed， that is the background knowledge was suppressed only at the high-level of the network， and the learning of basic graphic features was retained at the bottom of the network. Experimental results show that the student network trained by a specific knowledge learning algorithm can be as good as or even has better classification performance than a teacher network whose parameter scale is six times of that of the student network in specific category classification.

In the field of traffic safety， the road abandoned objects easily cause traffic accidents and become potential traffic safety hazards. Focusing on the problems of low recognition rate and poor detection effect for different abandoned objects of traditional road abandoned object detection methods， a road abandoned object detection algorithm based on the optimized instance segmentation model CenterMask was proposed. Firstly， the residual network ResNet50 optimized by dilated convolution was used as the backbone neural network to extract image features and carry out the multi-scale processing. Then， the Fully Convolutional One-Stage （FCOS） target detector optimized by Distance Intersection over Union （DIoU） function was used to realize the detection and classification of road abandoned objects. Finally， the spatial attention-guided mask was used as the mask segmentation branch to realize the object shape segmentation， and the model training was realized by the transfer learning method. Experimental results show that， the detection rate of the proposed algorithm for road abandoned objects is 94.82%， and compared with the common instance segmentation algorithm Mask Region-Convolutional Neural Network （Mask R-CNN）， the proposed road abandoned object detection algorithm has the Average Precision （AP） increased by 8.10 percentage points in bounding box detection.

Due to the lack of performance analysis while designing a distributed Evolutionary Algorithm (dEA), the designed algorithm cannot reach the expected speedup. To solve this problem, a comprehensive performance analysis method was proposed. According to the components of dEAs, factors that influence the performance of dEAs can be divided into three parts, namely, evolutionary cost, fitness evaluation cost and communication cost. Firstly, the feature of evolutionary cost under different individual encoding lengths was studied. Then when the evolutionary cost was kept unchanged, the fitness evaluation cost was controlled by using the delay function of the operating system and the communication cost was controlled by changing the length of individual encoding. Finally, the effect of each factor was tested through control variable method. The experimental results reveal the constraint relation among the three factors and point out the necessary conditions for speeding up dEAs.

The estimation of mixture matrix is a key step to solve the problem of blind source separation. But there lacks a general estimation method suitable for well-determined, over-determined and under-determined mixture matrix in the existing research. Scaling and permutation ambiguities lie in both factor matrix of tensor canonical decomposition and mixture matrix in blind source separation. With this property, the estimation of mixture matrix can be transformed into tensor canonical decomposition of observed signals' statistics. The canonical decomposition can be implemented by cyclic minimization, with the algorithm of alternating least squares. The theoretical analysis and simulations show that the method proposed in this paper is a general method to estimate well-determined, over-determined and under-determined mixture matrix.

A kinematic model of mobile robot with certain constraints of motion was established for unsmooth motion of three-wheeled mobile robot in the process of trajectory tracking control. According to the description of differential equation of mobile robot's position and orientation error, a trajectory tracking controller based on back stepping and time-varying state feedback was designed. It was proved that the controller could guarantee the uniformly asymptotical stability of the closed-loop system, according to the stability analysis of trajectory tracking controller. The simulation results verify the correctness of the kinematic model and the effectiveness of trajectory tracking controller.

Concerning the filtering problem of trees, buildings or other ground objects in field terrain reverse engineering, the disadvantages of conventional angle-limitation algorithm were analyzed, which accumulated errors or used a single threshold and could not meet the requirement of wavy terrain. Therefore, a self-adaptive angle-limitation algorithm based on scanning line was put forward. This method worked through limiting the angle of scanning center, reference point (known ground point) and the point to be sorted, which was adaptive with the wavy terrain. Then the modified point cloud was optimized with a curve fitting method by moving window. The experimental results prove that, the proposed algorithm has a sound control of the macro-terrain, and it can filter the wavy terrain point cloud much better.

As one of the core equipments in doubly-fed induction wind power generation system, the operation reliability of power converters seriously influences the safety and stability of power generation system. Since some flaws exist in Wavelet Neural Network (WNN) based on Recursive Least Square (RLS) algorithm such as low convergence precision and rate, and searching space possessing local minima and oscillation. The authors proposed a modified algorithm for fault detection of diagnostic power converters, in which variable weight and alter learning coefficient were employed to resolve above problems. After the modified WNN was trained and the faults were recognized from practical current data, comparison and analysis were carried out in simulation. The experimental results demonstrate that the modified algorithm can provide higher diagnostic precision and require less convergence time than the RLS algorithm.

Demanding on simultaneously carrying out the relations judgment of features′ elements containment as well as features′ geometry topotaxy in processing of GML path query, Tap the traditional XML path encode techniques analyzing, the paper designed an extended region code: ER-Code which takes the feature element as the unit of GML documents. Experimental results show that ERCode has good performances on initial construct and holistic query based on ER-Code. In addition, integrating ER-Code and the feature spatial geometry characteristic to an overall space can improve the efficiency that GML data path query greatly. This coding method has certain theory and practical value to the research of GML relevance technologies such as spatial data querying and storing.

A prototype system was presented which applied semantic Web services technology to automated business integration, focusing specifically on an automobile manufacturing supply chain. The system was able to handle order interaction, selection and generation through service discovery, selection and execution. Meanwhile this system adopted semantic expansion mechanism to enhance the capability of service discovery, search potential semantic Web services satisfying users requirements, and recommend services according to matching degree and preference. Besides, the architecture of the system was given, together with the implementation of key components.