Functional Integration and Interference in Embedded Control Systems

Project no: 9905-13

Martin Törngren and Jan Wikander

Royal Institute of Technology, Department of Machine Design, Division of Mechatronics (DAMEK)

Application,

Comments by evaluator 1, evaluator 2, ARTES board.

Comments and complements 990818, Support letter from SAAB Automobile AB

Support: 1 PhD student for 2 years decided 99-08-26.

Since 2000-12-01 is Jonas Norberg graduate student in this project.

Reports: Influence of ARTES support for one PhD student instead of two.

 

Industry contacts

Kenneth Lind kenneth.h.lind@saab.com

SAAB Automobile AB

Trollhättan

Project 1: Functional Integration and Interference in Embedded Control Systems

Project 2: Pre-Implementation Analysis of Distributed Control Systems - PICADOR

Overview

The automotive industry is facing an increasing extent of functionality that is implemented in software on interconnected electronic control units. When introducing control functions that are dynamically coupled through the controlled system, compare for example with active yaw control with brakes and active damping, it is not sufficient to simply superpose these controllers. Moreover, the new functions not only interact dynamically but also through more or less resource sharing, where a resource here refers to sensor(s), actuator(s) and (at least) the computer system. A very interesting aspect of this functional integration in future vehicles is that the systems will include a mixture of functions of different safety criticality levels. This includes both independent functions, coupled through implementation related resource sharing, and interrelated functions also being coupled dynamically. Unless appropriate measures are taken, each and every function involved in vehicle dynamics control would, since they rely on each other and interact, be required to be developed according to the most stringent of the SIL’s (risks). Thus, development would be very costly. Further, the dependence and integration of functions is expected to pose particular problems with respect to mode handling, and to include a mixture of functions with different timing characteristics and requirement.

To exploit the possibilities opened up by the resulting distributed computer infrastructures, this research project proposes research to investigate functional integration and interference from both a control functionality and computer implementation perspective. Interdisciplinary approaches in this area appear to be very rare in related research.

The following topics will be studied in the project:

• Control methods and architectures by which different control functions can be integrated in a cooperative manner to achieve a common goal, including methods for integration of and arbitration between control functions of different safety criticality.
• This new and extended functionality and global integration thereof requires new mechanisms and services from the real-time computation and communication subsystems, in particular to deal with interactions between functions of different criticality including aspects on interfacing, additional safety functions, and to deal with mode management including mode transitions due to functional interactions.

The innovativeness and uniqueness of this project is that it focuses on global systems integration in terms of integrating control functions which interact and may interfere with each other in terms of all the resources outlined above and in terms of dynamics of the rest of the controlled system.

Results and contributions

Co-operation

Publications