Project of Russian Foundation of Basic Research N 03-01-00323 (2003 - 2005):
Elaboration of multilevel hierarchical control systems on basis of method of autonomous adaptive control (AAC)
Abstract. The aim of the project is the development of the "Autonomous adaptive control" (AAC) method. The AAC method allows creating adaptive control systems on basis of a bionic approach. The AAC control systems are united recognition-controlling system, which perform learning and control in one process. On the stage of the AAC method development we want to elaborate multilevel and hierarchical structure of the AAC system. It will give us the increasing of the AAC system effectiveness and will extend the set of applications of the AAC systems. Possibilities of multilevel and hierarchical structure of the AAC system follow from the nature of the AAC method. The multilevel and hierarchical mode of the AAC system will be organized not via a complexness of the structure of the control system directly but via a complexness of the informative elements, which the AAC system works with - the "patterns" and "knowledge".
Simulation of process of control in cardio-vascular system by means of neuron-like Autonomous Adaptive Control (AAC) system
Abstract. The joint (ISP RAS and INRIA) project goal is investigation of functioning of human cardio-vascular system. We pay main attention to problem of the control in the cardio-vascular system, understanding that the control in human body is the privilege of special areas of nervous system. Idea of the project is to elaborate the simulation of controlled cardiovascular system by means of connection of two models - the mathematical model of cardiovascular system elaborated in INRIA and the simulation of nervous system ("Autonomous Adaptive Control system" (AAC)), elaborated in ISP RAS. We expect that the complex model will simulate important properties of cardiovascular system because the model will have following advantages - it will consist of a) the precise mathematical simulation of the cardiovascular system (as controlled object) and b) the neuron-like adaptive autonomous control system (AAC system). We plan to use the model for simulation of both the normal work of the cardiovascular system and some pathologic processes caused by abnormal work of nervous control system.
The development of the software tools suite "4GN" for supporting the process of modeling and prototyping or creation of applied AI systems on the basis of the Autonomous Adaptive Control (AAC) method
Abstract. The main objective of the project is the development of the software tools for the support of process prototyping of AI systems, based on AAC methodology. Implementation of the project should ensure the increase in quality of prototyping, reduction of time required for AI systems prototyping, and make the process of AI systems development more flexible and transferable. The purposes of the Project will be achieved by means of: unification of the AI system prototyping process, use of GUI (graphical user interface) for prototyping, use of facilities that shall allow integration of an AI system with external data sources, use of libraries of ready components of AI systems based on AAC method, library of developed and applied AI systems, library of various tests, automation of the documentation process.
Adaptive control system of transport robot
Abstract. The project aim is the creation of neuron-like adaptive control system on basis of AAC method, which simulates the nervous system. We use a mobile robot as a demonstrative and convenient object for the AAC control system. Moreover the mobile robotics is topical issue nowadays. We are creating the control system as a network of neuron-like elements. We have not a goal to develop mobile robot on any basis, we have the goal - to simulate self-learning neuron-like adaptive control system. However our mobile robot demonstrates new properties, for example it can train to bypass obstacles in real time of control without a "teacher". We use a program model of a mobile robot, which was named Gnome N8.
The information about some previous projects can be found in Background.