考虑驾驶员的多信息处理和多通道判断能力时车辆跟驰模型的改进

交通流中车辆跟驰建模的研究在交通工程和智能交通系统领域内显得十分重要 ,其核心就是对驾驶员认知问题的测度 .本文在总结以往跟驰模型的基础上具体讨论了 AP(action point)模型 ,并对其不足之处做了改进 ,重点研究了后车跟随加速度的算法 .此外 ,本文提出了驾驶员多信息处理机制和多通道判断决策概念 ,企图从控制论角度来对驾驶行为进行客观地定性分析和定量描述 ,以期正确地揭示交通流的本质特征 .

生态环境保护区域的地质灾害气象风险预警系统

为了能够及时预测生态环境保护区域的地质灾害情况、减少地质灾害造成的损失,基于多信息融合过程,设计了一种地质灾害气象风险预警系统。在具体设计前,针对生态环境保护区域展开分析,探讨地质灾害造成的影响;在系统硬件部分,重点设计了处理器、通信模块、终端显示模块与加速度传感器模块;在系统软件部分,收集地质灾害资料、展示与维护地质灾害资料,利用多信息融合技术对地质灾害气象风险信息实施融合处理,根据D-S证据理论的组合准则,求出各证据的联合概率分布函数,从而实现对地质灾害气象风险的预警。实例分析表明,该系统能够有效对降雨量与灾害等级作出预警,为生态环境保护区域的防灾减灾作出指导,满足设计预期。

Multi-sensor systems and information processing of mobile robot in uncertain environments

The PBJ- 01 robot is a kind of mobile robot featuring six wheels and two swing arms which can help it to fit many terrains. The robot has a sophisticated sensor system, which includes ultrasonic sensors, tentacle sensors and a vision sensor. The PBJ- 01 adopts behavior-based reactive control architecture in which the key part is an object recognition system based on a fuzzy neural network. Simulation validates that this system can conclude the obstacle type from the sensor data, and help the robot decide whether to negotiate or to avoid obstacles.

Radar target recognition based on micro-Doppler effect

Mechanical vibration of target structures will modulate the phase function of radar backscattering,and will induce the frequency modulation of returned signals from the target. It generates a side bands of the target body Doppler frequency shift,which is helpful for target recognition. Based on this,a micro-Doppler atomic storehouse is built for the target recognition,and four kinds of common classifiers are used separately to perform the classified recognition. The simulation experimental results show that this method has high recognition rate above 90%.

Understanding visual-auditory correlation from heterogeneous features for cross-media retrieval

Cross-media retrieval is an interesting research topic,which seeks to remove the barriers among different modalities.To enable cross-media retrieval,it is needed to find the correlation measures between heterogeneous low-level features and to judge the semantic similarity.This paper presents a novel approach to learn cross-media correlation between visual features and auditory features for image-audio retrieval.A semi-supervised correlation preserving mapping(SSCPM)method is described to construct the isomorphic SSCPM subspace where canonical correlations between the original visual and auditory features are further preserved.Subspace optimization algorithm is proposed to improve the local image cluster and audio cluster quality in an interactive way.A unique relevance feedback strategy is developed to update the knowledge of cross-media correlation by learning from user behaviors,so retrieval performance is enhanced in a progressive manner.Experimental results show that the performance of our approach is effective.

Evaluation of Multistrategy Classifiers for Heterogeneous Ontology Matching On the Semantic Web

On the semantic web, data interoperability and ontology heterogeneity are becoming ever more important issues. To resolve these problems, multiple classification methods can be used to learn the matching between ontologies. The paper uses the general statistic classification method to discover category features in data instances and use the first-order learning algorithm FOIL to exploit the semantic relations among data instances. When using multistrategy learning approach, a central problem is the evaluation of multistrategy classifiers. The goal and the conditions of using multistrategy classifiers within ontology matching are different from the ones for general text classification. This paper describes the combination rule of multiple classifiers called the Best Outstanding Champion, which is suitable for heterogeneous ontology mapping. On the prediction results of individual methods, the method can well accumulate the correct matching of alone classifier. The experiments show that the approach achieves high accuracy on real-world domain.

Research on multi-document summarization based on latent semantic indexing

A multi-document summarization method based on Latent Semantic Indexing (LSI) is proposed. The method combines several reports on the same issue into a matrix of terms and sentences, and uses a Singular Value Decomposition (SVD) to reduce the dimension of the matrix and extract features, and then the sentence similarity is computed. The sentences are clustered according to similarity of sentences. The centroid sentences are selected from each class. Finally, the selected sentences are ordered to generate the summarization. The evaluation and results are presented, which prove that the proposed methods are efficient.

Design and implementation of a large-scale multi-class text classifier

Although, researchers in the ATC field have done a wide range of work based on SVM, almost all existing approaches utilize an empirical model of selection algorithms. Their attempts to model automatic selection in practical, large-scale, text classification systems have been limited. In this paper, we propose a new model selection algorithm that utilizes the DDAG learning architecture. This architecture derives a new large-scale text classifier with very good performance. Experimental results show that the proposed algorithm has good efficiency and the necessary generalization capability while handling large-scale multi-class text classification tasks.

Experimental verification of genuine multipartite entanglement without shared reference frames

Quantum entanglement is an essential resource for quantum information processing, either for quantum communication or for quantum computation. The multi- partite case of entanglement, especially the so called gen- uine multipartite entanglement, has significant importance for multipartite quantum information protocols. Thus, it is a natural requirement to experimentally verify multipartite quantum entanglement when performing many quantum int^rmation tasks. However, this is often technically chal- lenging due to the difficulty of building a shared reference lYame among all involved parties, especially when these parties are distant l^om each other. In this paper, we experimentally verify the genuine tripartite entanglement of a three-photon Greenberger-Horne-Zeilinger state without shared reference frames. A high probability 0.79 of successfully verifying the genuine tripartite entanglement is achieved when no reference frame is shared. In the case of sharing only one common axis, an even higher success probability of 0.91 is achieved.

Coherent state transfer through a multi-channel quantum network： Natural versus controlled evolution passage

Quantum state transfer(QST) is an important task in quantum information processing. In this study, we describe two approaches for the high-fidelity transfer of a quantum state between two opposite quantum dots attached to a multi-channel quantum network.First, we demonstrate that a high-efficiency QST can be achieved with the coherent time evolution of a quantum system without any external control. Second, we present an approach that uses an alternative mechanism for a high-fidelity QST. By adiabatically varying tunnel couplings, it is possible to implement the complete transmission of a quantum state based on this quantum mechanical mechanism.