Design of adaptive, multidimensional information and control systems

当今计算系统的数据收集（遥感、传感器网络等）、处理和联网能力的急剧增加，意味着现在可以访问大量数据，并在设计和控制多样化、通常是复杂的系统时有效地使用它们。这些发展对智能信息和控制系统领域的发展产生了深远的影响。在拟议的研究计划中，主要目标是开发一种综合方法，解决在动态、不确定环境中大量数据的分布式处理背景下设计和控制自适应、自组织系统中的这些问题。具体的长期目标包括：-发展系统设计的理论和方法基础，根据功能和性能要求结合不同类型的信息处理技术来调整其结构。-处理因处理大量多维信息而产生的要求（降维、性能标准等）。——发展以神经网络、模糊逻辑、精选统计技术和自适应索引为主的混合技术，并将其应用于对适应性和学习能力要求最高的非线性、时变和不确定系统。本研究项目的成果，在促进控制系统和模式识别领域理论发展的同时，将在遥感、地理信息学、智能制造系统、分布式测控系统和决策支持系统等多个领域得到应用。

The sharp increase in data gathering (remote sensing, sensor networks, etc.), processing and networking capabilities of today's computing systems means that it is now possible to access vast amounts of data and to efficiently use them in the design and control of diversified, often complex systems. These developments have had a profound impact on the evolution of the area of intelligent information and control systems. In the proposed research program, the main objective is the development of an integrated methodology that addresses those issues in the design and control of adaptive, self-organizing systems that are critical in the context of distributed processing of large quantities of data in dynamic, uncertain environments. Specific long-term objectives include: - Development of theoretical and methodological foundations for the design of systems that adapt their structure by combining different types of information processing technique according to functional and performance requirements. - Addressing the requirements imposed by the fact of dealing with large amounts of multi-dimensional information (dimensionality reduction, performance criteria, etc.).   - Development of hybrid technologies integrating mainly neural networks, fuzzy logic, selected statistical techniques and adaptive indexing, and their application in non-linear, time-variable, uncertain systems, where the need for adaptability and learning capabilities is most pronounced. The results of this research program, while contributing to theoretical developments in the areas of control systems and pattern recognition, will at the same time be applicable in several diversified domains, such as remote sensing, geomatics, Intelligent Manufacturing Systems, distributed measurement and control systems, and decision support systems.