Visible to the public File preview

London Workshop on the Control of Cyber-Physical Systems

… On the (new) discipline: Systems and Control Science

From the HYCON2 position paper (copies in this room)

Workshop on CPS, London, October 20-21, 2012

1

Control is ubiquitous

• No sophisticated device (mobile phone, camera, washing machine, car, train, …) without lots of control loops • High-precision and energy efficient manufacturing require precise control of a large number of variables • Control technology constitutes a large part of the turnover of the European electrical/ electronics industry High-performance control is a key technology for European competitiveness

Workshop on CPS, London, October 20-21, 2012

2

Systems and control science – what is it about? Systems and Control Science provides the foundation for the realization of control and real-time decision support systems:
 Mathematical modelling of physical and social phenomena to understand and to predict their dynamic behaviours and the interactions with other systems  Design of control strategies and algorithms to optimize the behaviour of systems so that they accomplish the intended functions and satisfy constraints, and such that negative effects are minimized

 Implementation of control strategies by sensing devices, actuators, computing elements, communication systems,... and integrating them into a performing system under cost constraints
 Validation and verification that the implemented control strategies together with the physical or social systems, satisfy constraints and performance requirements.
Workshop on CPS, London, October 20-21, 2012
3

Control: Generic contributions and applications
Close BI-DIRECTIONAL interaction between methodological developments & advances in applications
Control theory Modeling Optimization Signal processing Simulation

METHODS

Verification Definition of new abstract problems Biology

Machine Learning
Modeling and control tools Energy

 abstract mathematical representations  general solutions to theoretical problems  general tools for analysis and design

Water Automotive
Chemical

Health
Robotics Aerospace

SYSTEMS
Transportation

Manufacturing

Economics

Finance

Environment

 specific performance requirements  specific modeling needs  specific HW/SW implementation contraints constraints  tailored solutions to complex problems
4

Workshop on CPS, London, October 20-21, 2012

SYSTEMS (Complex, SoS, CPS, etc…) - Different layers: each layer has its own *life* (size and time scale of models) but it is strongly connected to the other layers: small variations at a particular point in the lowest layer can generate chaos or blackout at the highest level New fondamental approaches and tools are needed

higher level: networks of interactions

Systems and Control Science provided ,and is providing very efficient methods and technologies

intermediate level: interactions between systems

lower level: devices to control

Workshop on CPS, London, October 20-21, 2012

5

Ground traffic and air traffic management

Ground traffic and air traffic management are largescale control problems with distributed decision makers. 3) traffic flow 2) interactions between electronic devices 1) electronic devices for making vehicles cleaner or safer

Workshop on CPS, London, October 20-21, 2012

6

3)

2) 1)

Workshop on CPS, London, October 20-21, 2012

7

Contribution of control to societal needs

3)

2)

1) 3) overall network security, mitigation of large disturbances 2) balancing produced and consumed power 1) electro-technical devices for ensuring local distribution
Workshop on CPS, London, October 20-21, 2012
8

Cellular and biomolecular research

3) networks of interactions (transcription networks) 2) interactions between molecular species 1) microfluidic devices to control the cellular environment

Workshop on CPS, London, October 20-21, 2012

9

Key Challenges
… On these three levels we need:
• Make full use of the embedded computing power in all technical systems to achieve optimal performance • Extract useful information from the enormous amount of data acquired, communicated, and stored in technical systems • Construct dependable, performing, evolvable and evolving systems with reduced engineering effort. • Guarantee safety and user-friendliness of automated systems • …. CONTROL …

Workshop on CPS, London, October 20-21, 2012

10

Five key challenges for future R&D and Innovation

1. System-wide coordination and control of large-scale systems

CPS & Systems and Control Science

2. Distributed networked dynamic systems

HORIZON HORIZON 2020 2020

3. Autonomy, cognition and control 4. Model-based systems engineering
5. Human-machine interaction in automated systems

Workshop on CPS, London, October 20-21, 2012

11

Summary

• Advanced control technology is a key for
• the competitiveness of the European manufacturing industry • high performance consumer products with embedded computing with which Europe has success on the world market • the satisfaction of urgent societal (health, economics, finance, environment) needs

• The strength of Systems and Control Science is its applicability across all industrial and societal sectors and that it transfers developments from one sector to the others, thus leading to innovations on a broad scale.

Workshop on CPS, London, October 20-21, 2012

12

Systems and Control Science (good key word)
= NEW DISCIPLINE merging Control, Communication and Computer Science at the heart of the future technologies for “PHYSICAL” and “SOCIAL” systems

Workshop on CPS, London, October 20-21, 2012

13

System-wide coordination and control

• Enormous needs, chances and challenges for economic and ecological optimization in the management of complex systems • Deep vertical integration and standardized communication needed!
Workshop on CPS, London, October 20-21, 2012
14

Distributed and networked dynamic systems
Wireless communications are revolutionizing the way factories, buildings, transportation systems (cars, airplanes, trains and ships), and homes are automated. New standards for wireless communication are needed for feedback control purposes
 Control

methods for networked systems must take into consideration that communication links may be interrupted or be established only for limited periods of time.
between subsystems is constrained is a new challenge that has to be tackled by the control and the communication community together.

 Ensuring stability and performance for systems in which communication

Workshop on CPS, London, October 20-21, 2012

15

Autonomy, cognition, and control
Systems with (partial) autonomy
• Exhibit

goal-oriented behaviour and flexibly change their goals and behaviour depending on context and experience • Can act in unstructured environments without human intervention and robustly respond to dynamic changes of the environment • Can Interact with humans and with other cognitive systems to jointly solve a task. • Will be increasingly important for assistance to the elderly in an ageing society!

How to ensure safe and certified operation for highly autonomous systems under all conceivable conditions? How to combine data-based techniques (machine learning) with rigorously founded algorithms to achieve unprecedented performance? How to extract valuable information from large data sets?
Workshop on CPS, London, October 20-21, 2012
16

Model-based systems engineering
• Efficient and reliable systems design requires the

use of models. • Different aspects of technical systems are often modeled by different formalisms that are the basis of the use of specific tools. • This calls for the integration of different formalisms into one environment with efficient re-use of models and integration of results obtained by different tools.

• Integration of mathematically sound control design and analysis methods into industrial environments where heterogeneous data, documentation, models, formalisms and software are used on all design levels with presently weak coupling and no guarantee for consistency. • Advanced methods for automated design flow support, such as consistency checking and dependency tracking Model-based design is the key to the development of dependable complex systems and of systems of systems.
Workshop on CPS, London, October 20-21, 2012
17

Human-machine interaction in automated systems
• The interplay of humans with highly automated

systems with complex dynamics and complex controllers is crucial for the functioning and performance of large technical systems. • This area needs a much deeper investigation, bringing together researchers from cognition, human-machine interaction and systems and control. Challenge: Development of a synergistic approach where the capabilities of complex and theoretically well-founded control systems and of humans are combined in the most effective manner to achieve the best possible system performance while guaranteeing safety and integrity.

Requires collaborations of control scientists with behavioural science.

Workshop on CPS, London, October 20-21, 2012

18

Feedback control is the key for the future

“Real-time measurement, modelling and control platforms will drive a smarter planet through the broad implementation of feedback control.”
Dario Gill, Program Director, Energy Technology and Strategy, IBM T.J. Watson Research Center

Control algorithms are the core of most systems!

Workshop on CPS, London, October 20-21, 2012

19