i-Sense Activities


The project activities are organized in the following Work packages:

  • WP1 System Specification and Architecture
  • WP2 Cognitive Fault Diagnosis
  • WP3 Adaptation and Learning
  • WP4 Fault Tolerant Control
  • WP5 Integration, Validation and Evaluation of the iSense Platform
  • WP6 Dissemination
  • WP7 Management & Coordination


The figure below presents the organization of the Work packages in activities.


The aims of the i-Sense project will be achieved through the following scientific and technological objectives:

1. To develop a rigorous formulation for cognitive fault diagnosis and fault tolerant control problems
The developed formulation for cognitive fault diagnosis and fault tolerant control will specify the system characteristics to facilitate the development of a foundation for designing a new class of fault diagnosis algorithms. A categorization of possible faults models will be formulated, both in terms of their functional properties, and also in terms of their time evolution. A class of system architectures for cognitive fault diagnosis and fault tolerant control will be defined, which will facilitate the design and implementation of the iSense Platform.

2. To design cognitive fault diagnosis schemes that can be effectively applied to monitoring and control applications of uncertain distributed environments
Cognitive fault diagnosis is a novel approach for autonomous fault diagnosis that employs cognition and machine learning methods to enhance the fault diagnosis performance and make the overall system more adaptive and robust. The use of cognitive fault diagnosis approaches becomes more crucial as engineering systems are required to perceive and interact with more unstructured, open-ended environments. The developed framework for cognitive fault diagnosis will be achieved through the design of fault detection, fault isolation and fault identification schemes. Finally, the performance properties of the designed cognitive fault diagnosis schemes will be analyzed.

3. To develop a set of adaptation and learning algorithms that can be incorporated into the cognitive fault diagnosis and fault tolerant control schemes
The use of adaptation and learning aims to discover and exploit spatial-temporal relations that exist in collected data. These relations, which will be refined online during operation of the intelligent sensing system, will enhance robustness and tolerance to fault events and other unexpected field situations. Neural network ensemble learning, adaptive classification methods and virtual sensor/actuator schemes will be developed in the context of cognitive fault diagnosis.

4. To investigate the design, analysis and evaluation of fault-tolerant control schemes
While fault detection and identification determine the presence and nature of a fault, ultimately it is of fundamental importance to be able to use this knowledge for handling various fault scenarios and avoiding fault propagation. Fault-tolerant control refers to feedback control systems that are designed to withstand a certain measure of fault functions by sufficiently safeguarding the key operation of the system. Both passive and active fault tolerant control architectures will be examined and the stability and convergence properties of the developed fault tolerant control schemes will be pursued.

5. To integrate the various components and build a system prototype for the iSense Platform for intelligent building applications
The performance of the iSense Platform will be validated in the context of intelligent building application domain under a wide range of fault scenarios. The expected enhancement in fault tolerance achieved by the use of the iSense Platform will be evaluated, and possible fundamental limitations in achieving cognitive fault diagnosis and fault tolerant control will be identified. Finally, existing standards for building automation systems will be examined and contributions will be made in these standards to support fault tolerance.






Project Newsletter

Third issue January 2014

Second issue January 2013

First issue January 2012

Subscribe to the newsletter »