预警装备体系的发展和作战运用离不开反导预警系统作战效能评估,在体系作战效能层面上,以完成反导作战任务的目的作为牵引,对反导预警系统进行数学建模和量化评估.在对反导预警作战流程分析的基础上,建立反导预警系统作战效能评估指标体系,采用基于层次分析法和熵权法综合赋权的方法确定各指标的最终权重,利用改进的TOPSIS法对系统的作战效能进行了定量分析.通过对预警装备部署方案的仿真试验,证明了所提出的模型和算法的正确性和有效性,结果表明,相比于单纯使用熵权法确定权重,利用组合赋权法分析作战能力可信度更高.

To reuse existing motion data and generate new motion, a method of human motion nonlinear dimensionality reduction and generation, based on fast adaptive scaled Gaussian process latent variable models, is proposed. Through statistical learning on motion data, the motion data are mapped from high-dimensional observation space to low-dimensional latent space to implement nonlinear dimensionality reduction, and probability distributing of posture space which measures the nature of posture is obtained. The posture which meets constraints and has maximal probability can be computed as the solution of inverse kinematics. This method can avoid cockamamie computation and posture distortion existing in traditional inverse kinematics. The experiments show that our method has higher convergence velocity and precision and extends editing range of motion by adapting motion editing direction.

Present a technique of human motion extension to reuse existing motion capture data,which contains lengthways extension and transverse extension.Lengthways extension extends motion along time axis,generating new motion containing more frames;transverse extension extends motion along characters,which generates group motion,similar but not same each other,from single motion.Both lengthways extension and transverse extension are based on dynamic model for motion which describes the changing rule of motion posture from the angle of probability.Motion extension can reuse existing motion capture data to make motion generation simpler and extend application range of existing motion capture data.Experiments show that our method can generate customized fames amount motion and group motion from existing motion.

By studying the problems of parallel isosurface extraction and rendering,a novel framework is proposed according to the characters of 3D scalar fields.This framework adopts the static task distribution method to keep balance and Marching Tetrahedra algorithm to extract isosurface,and proposes a mixed parallel rendering and combining method based on rendering node control model,which is prearranged for parallel rendering and scene integration.An effective improvement on the capability of parallel isosurface extraction and rendering of large-scale 3D scalar fields is given by the experiments.

After researching the load balancing problem in parallel rendering,an adaptive load balancing algorithm for volume rendering is proposed according to the character of volume data.This method is an improved Whiteman algorithm,it calculates the depth of four vertexes of grid to set the grid's value and used time statistic to control the occasion of redistribution.Experiments show that this method can effectively improve the stability and capability of parallel volume rendering.

Parallel Rendering is an important fast visualization method of large-scale 3D scalar field. Two classes of rendering methods of 3D scalar field were depicted： surface rendering and direct volume rendering in the first and the parallel model, load balance and image compositing algorithms in parallel rendering were compared based on those two visualization methods. Then the existing parallel systems were classified and compared. At the last, some new problems and ideas were proposed for the next research after summarizing the existing algorithms.

This paper presented the concept of human lower limbs vector,which could retain the primary feature of motion.After that,it presented a method of human motion retargeting based on fixity of lower limbs vector feature to reuse the existing motion capture data better.This motion retargeting method oriented human lower limbs,which retargeted the motion data from original human skeleton model to target human skeleton model which had different bone length.The retargeted motion retained the primary feature of original motion.The results of experiment show that this method has good motion retargeting effect and high computing efficiency.

An inverse kinematics solution based on statistic learning is presented.Because of the high dimension of character animation motion data and correlation lying in various dimensions,it is a very hard work to analyze and compute directly.The motion data are mapped from high-dimensional space to two-dimensional latent space based on Gaussian process latent variable models(GPLVM).Then,the representative poses of virtual characters are found out by clustering the motion data in latent space,which can expand a subspace that contains the primary characters and disciplinarians of training data.Finally,the weight of representative poses is optimized combined with constraints on the end effectors,and the optimized pose is obtained.The experiments show that the proposed method obtains a better effect.

After researching the image compositing problem in sort-last parallel rendering, A dynamic image compositing method based on pre-frame time distributing is proposed. This method is a improved method of Direct-Send method. It considers the rendering node's capability as the important factor in image compositing. It used pre-frame time distributing to decompose the screen before image compositing. It adopt space instead of time method to balance the rendering nodes. Experiment show that this method can effectively improve the capability of sort-last parallel rendering.

Too much computation is the bottleneck in analysis of airborne antenna radiation.A parallel computing and rendering method is proposed to solve this problem.It proposes a semi-automatic algorithm to distill model's skeleton based on space octree dividing and a equal triangles dividing method to solve the load balancing problem.It adopts sort-last parallel rendering framework and Binary-swap image composition method to solve parallel rendering problem.The algorithm is implemented in PC cluster system.Experimental results show this algorithm can drastically reduce the computing time and improve the rendering effect.It is benefit to apply this method in analysis of airborne antenna radiation.

After researching the framework building of parallel computing and rendering and according to the character of radar net's detection ability,a framework of parallel computing and rendering based on PC cluster is proposed.In this framework,an octree is adopted to distribute the space of radar net in parallel computing,and a sort-last parallel mode is adopted in parallel rendering.The experiments show that this framework can satisfy the requirement of a large scale radar net effectively.

Propose a motion prediction technique based on Gaussian process dynamical models, which maps the existing motion to a low-dimensional latent space by nonlinear method and models the dynamics of motion with Markov chain in latent space. After this, a smooth latent trajectory is obtained corresponding to the motion. Then, predict the future latent states follow the tail end of latent trajectory to obtain a new one. Finally, map this new latent trajectory back to observation space to implement motion prediction. The new motion frames by prediction follow the motion rule of existing motion. Experiment shows that our method can automatically synthesize longer motion by prediction with existing motion.

This paper proposed an automatic segmentation technique for motion capture data.It reduced the dimension of motion capture data with Gaussian process latent variable models,mapped the motion capture data from high-dimensional observation space to low-dimensional latent space.Construced motion character function in latent space,which had much excellence such as simple construction,sensitive to all joints,and so on.By analyzing geometry character of motion character function,it could detect the segmentation point in motion capture data,and segment the motion.Experiments show that this technique has high correct rate and well adaptation.

This paper presents a novel method for retargeting motions. In this method we consider the whole leg as a length changeable skeleton, through keeping the length proportion and direction of the leg vector before and after retargeting we can accomplish the motion retargeting by scale the root node. Because we transform the constraint of foot position to the constraints of the leg vector's length and direction and adjusting the leg vector is easy, our method need not the complex optimization algorithm. The experimental results show that the method is a real-time method and the characteristic of foot trajectory can be kept after retargeting.

In order to void cockamamie computation and pose distortion existing in traditional inverse kinematics,this paper presents a Fast Adaptive Scaled Gaussian Process Latent Variable Model(FASGPLVM),then realizes human motion generation based on it.Experimental results show that FASGPLVM has higher convergence velocity and precision and extends editing range of motion capture data by adapting motion editing direction.

A parallel rendering method of large-scale battlefield electromagnetism based on PC cluster is proposed.This meth-od adopts ray casting direct volume rendering algorithm to render the electromagnetic environment,uses Octree to divide the rendering space and adopts preorder traversal Octree method to distribute the divided space.At last,Binary-swap algorithm is used to compose the images which are produced by PC nodes of PC cluster.Experiments show that the method can effectively satisfy the real time rendering of large-scale battlefield electromagnetic environment.

A group animation generation method based on machine learning was proposed in order to reduce the complexity of generating mass of similar but different natural human motions in group animations. There are two models. Poses learning model was built based on Gaussian process latent variable model to characterize a specific motion and dynamic model was built in latent space to characterize the dynamic evolving process of neighboring poses in latent space. These models can be represented as probability distribution over all poses composing the motion by learning from existing motion data. Dynamic prediction can be made in latent space for giving initial state,then hundreds of latent trajectories by Hybrid Monte Carlo sampling according to given probability distribution can be obtained. Group animations can be implemented by generating a series of similar but different natural motions reconstructed from these latent trajectories,thereby avoid the difficulty and complexity of calculating geometric relationship and physical constrains in inverse kinematics.

We present an inverse kinematics implementation technique based on statistical learning. Because of the high dimension of character animation motion data, direct analysis on them is a very hard work. We map the motion data from high-dimensional observing space to two-dimensional latent space, based on Gaussian process latent variable models (GP-LVM), then, find out the representative poses of virtual character by clustering the motion data in latent space. Finally, weight the representative poses and optimize the weights, combined with constraints on the end effectors, and synthesize the optimized pose. The experiments show that our method obtains satisfying effect.

This paper studies the designing and building thought of GeoFusion platform, uses other digital earth platforms for reference, analyzes different kinds of objective factors in battlefield environment, visualizes the battlefield situation information, and simulates common operation picture according to the requirement of the officer. In the demonstration process of the system, the officer may choose the observation way and object, adjust the angle of view and the viewing position. The system has strong real-time interaction.