People

Research activity in 2000

Model based reasoning has been the main area of investigation of our group in recent years. Several aspects have been explored.

From a methodological point of view, the recent contributions can be grouped as follows:

• Analysis of the problems arising when modelling and diagnosing complex dynamic systems. We paid specific attention to the study of the different dimensions that have to be taken into account when dealing with dynamic, time dependent and time varying behaviour and to the approaches that can be used to face these problems.
• Modelling with Process algebras. We investigated the use of the formalism of Process algebras (which is widely used for modelling in other areas of computer science) which proved to be a very flexible and powerful language for model-based reasoning.
• Computational issues in the diagnosis of dynamic systems. Diagnosing these systems is very complex and thus focusing techniques have to be adopted. We worked on the extension of techniques developed in the past for static systems (knowledge compilation, use of causal knowledge to focus the simulation of dynamic behaviour).
• Embedding diagnostic systems into technical devices. This part of the work is the one most closely related to application and focused on the problem of generating, starting from a model-based approach, the diagnostic rules that can be implemented in the Electronic Control Units of cars (more generally, the diagnostic rules that can be embedded in electronic system that control complex devices). We defined a new approach to the design and construction of diagnostic decision trees (fault trees) which include the notions of time and information about control actions, costs of these actions, temporal constraints for control.
• Computational issues in multi-modal approach to diagnosis. In the last few years we have investigated how to integrate case-based reasoning with model based reasoning for solving diagnostic problems. We have shown that significant gains in terms of computation efforts can be gained when it is possible to decide to activate the CBR or the MBR component on the basis of the difficulty of the diagnostic problem at hand. In particular we have devised an heuristic method able to estimate the difficulty (in terms of computational effort) of a diagnostic problem. We have also investigated which is the impact of the content of the case memory on the quality of the solution of a diagnostic problem as well as on the performance of the multi-modal reasoning system. Learning techniques able to decide when to add a new case to the case memory and when cases have to be forgotten have been developed.

The main area of application is the automotive one. After the experience of the BRITE-EURAM "Vehicle Model-Based Diagnosis" project, in which we developed diagnostic software that can run on board cars (the project has been demonstrated on the "Common Rail" system for diesel engine on a prototype car, a Lancia Kappa equipped with a 2.4 liters Common Rail diesel engine), we started a new European Project: The IDD (Integrating Diagnosis and Design).

The project is supported by the EU under the Substainable Growth Programme (V Framework) and the partners are: Centro Ricerche Fiat, Daimler Chrysler, Renault, PSA - Peugeout Citroen, Magneti Marelli, Occ’m, Universitè Paris XIII, Technical University Munchen, Università di Torino.

The aim of the project is to integrate model-based diagnostic techniques in the tools used during the design of a system (CAD tools, simulation tools such as Matlab/Simulink). In this way the diagnosability of a system can be verified (and improved) during he design process and the diagnostic rules to be implemented in the control of the system can be produced automatically at the end of the design.

Model-based reasoning has been applied to another relevant domain for diagnostic problem solving. In a cooperative project supported by ASI (Agenzia Spaziale Italiana) we have developed the diagnostic module. In particular, we have investigated how to represent knowledge at different degrees of abstraction and we propose an innovative method for representing diagnoses in a concise way. In particular, the mechanism is based on a formulation of diagnosis in terms of constraints and variable assignment. Such a representation allows one to develop efficient methods also for diagnosis generation. The domain chosen as test-bed is the space robotic arm SPIDER and a prototype of the diagnostic module for such a device has been developed.