Francesco Flammini discusses the importance of digital twins for enhancing the resilience, safety, and autonomy of cyber-physical systems (CPS) in his work. He emphasizes the need for advanced modeling techniques to handle complexity and uncertainty and outlines a framework for the dynamic management of CPS through self-healing and predictive capabilities. The document covers topics like model-based engineering, risk factors, and the integration of AI in ensuring trustworthy and safe systems.

1. Francesco Flammini
Digital Twins for Trustworthy Autonomy
Francesco Flammini francesco.flammini@ieee.org

2. Francesco Flammini
Smart-X applications and the Internet of everything

3. Francesco Flammini
GRAND CHALLENGES
Need for designing cyber-physical systems (CPS) that are:
- Ubiquitous and pervasive
- Smart, intelligent and autonomous
- Reliable, safe and secure
How to manage complexity, heterogeneity and uncertainties?
F. Flammini et al. (2019). Complex, Resilient and Smart Systems. In: (edited by): F. Flammini, Resilience of Cyber-Physical
Systems: From Risk Modeling to Threat Counteraction. ADVANCED SCIENCES AND TECHNOLOGIES FOR SECURITY
APPLICATIONS, p. 3-24, BERLIN HEIDELBERG:Springer-Verlag, ISBN 978-3-319-95597-1, ISSN: 1613-5113

4. Francesco Flammini
What is resilience?
“The persistence of service delivery that can justifiably be trusted, when facing changes.”
“The persistence of the avoidance of failures that are unacceptably frequent or severe,
when facing changes.”
“The persistence of dependability when facing changes.”
(Jean-Claude Laprie, 2008)

5. Francesco Flammini
What is RISK?
R = P x V x D
Several quantitative non-functional requirements, e.g.:
MTBF > 500.000 h
THR < 10-9 hazards / h

6. Francesco Flammini
Model-based engineering
https://modeling-languages.com/clarifying-concepts-mbe-vs-mde-vs-mdd-vs-mda/

7. Francesco Flammini
Multi-paradigm modeling
OP3
(Generalized
Stochastic)
Petri Net
OFF SB SR
Level 0 /
STM
System Failure
Full Supervision
Level 1 Level 2
O
p
e
r
a
t
i
n
g
M
o
d
e
s
L
a
y
e
r
H
a
r
d
w
a
r
e
L
a
y
e
r
P
r
o
c
e
d
u
r
e
s
L
a
y
e
r
Start of Mission Hand-Over
Level 3
Unfitted
OS
OP2
fail SS 1
fail SS 2
fail 11
fail 3
fail 4
fail 2
System Failure
v4
v5
v2
v1
v3
OP1
Finite State Machine
Finite State Machine
(Repairable) Fault Tree
Bayesian Network
(Generalized
Stochastic)
Petri Net
OP3
(Generalized
Stochastic)
Petri Net
I
n
t
e
r
m
e
d
i
a
t
e
S
o
f
t
w
a
r
e
L
a
y
e
r
s
H
a
r
d
w
a
r
e
L
a
y
e
r
P
r
o
c
e
d
u
r
e
s
L
a
y
e
r
Start of Mission Train 1 Hand-Over Train 2
(Generalized
Stochastic)
Petri Net
ON-BOARD SUBSYSTEM 1
TRACKSIDE SUBSYSTEM
OP3
(Generalized
Stochastic)
Petri Net
O
p
e
r
a
t
i
n
g
M
o
d
e
s
L
a
y
e
r
H
a
r
d
w
a
r
e
L
a
y
e
r
P
r
o
c
e
d
u
r
e
s
L
a
y
e
r
Start of Mission Hand-Over
(Generalized
Stochastic)
Petri Net
ON-BOARD SUBSYSTEM n
S
o
f
t
w
a
r
e
L
a
y
e
r
H
a
r
d
w
a
r
e
L
a
y
e
r
LINESIDE SUBSYSTEM
...
OPi 1
Non
Transmitting
Transmitting
Correct T elegram
Transmitting Default
Telegram (safe failure)
Transmitting Uncorrect
Telegram (unsafe failure)
BALISE 1
BALISE K
...
GROUND SUBSYSTEM
OPi 2
A Study on Multiformalism Modeling of Critical Infrastructures.
3rd International Workshop on Critical Information Infrastructures
Security, CRITIS’08, Frascati (Rome), Italy, October 13-15, 2008

8. Francesco Flammini
Certification

9. Francesco Flammini

10. Francesco Flammini
Self-x technologies
N. Rajabli, F. Flammini, R. Nardone and V. Vittorini, "Software Verification and
Validation of Safe Autonomous Cars: A Systematic Literature Review," in IEEE
Access, vol. 9, pp. 4797-4819, 2021, doi: 10.1109/ACCESS.2020.3048047.

11. Francesco Flammini
Adversarial attacks to AI

12. Francesco Flammini
AI for safety critical systems
▸ https://www.eetimes.eu/2019/02/20/ai-tradeoff-accuracy-or-robustness/
▸ https://dorsa.fyi/cs521/
(IBM Research)

13. Francesco Flammini
Trustworthy artificial intelligence
https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai

14. Francesco Flammini
Explainable artificial intelligence
https://www.darpa.mil/program/explainable-artificial-intelligence

15. Francesco Flammini
Self-adaptive systems
IBM Corporation. An architectural blueprint for autonomic computing (April 2003), https://www-03.ibm.com/autonomic/pdfs/AC%20Blueprint%20White%20Paper%20V7.pdf

16. Francesco Flammini
Resilience through smart-troubleshooting
Caporuscio, M., Flammini, F., Khakpour, N., Singh, P., and Thornadtsson, J.
(2020) “Smart Troubleshooting Connected Devices: Concepts, Challenges
and Opportunities”. Future Generation Computer Systems (Elsevier)

17. Francesco Flammini
Self-healing

18. Francesco Flammini
Digital Twins of Cyber-Physical Systems

19. Francesco Flammini
Digital Twins as Run-Time Predictive Models
Start
Dep. target
achieved?
Shut-down and
switchover to
fallback procedure
or redundancy
Stop
No
Yes
Build reference
DT model
Instantiate DT
model with initial
configuration data
Evaluate DT model
to estimate system
dependability
Update DT model
with new plan/
configuration
Compute plans/
reconfiguration
options
Acceptable
options exist?
Fault or
anomaly?
No
Yes
Yes
No
Select optimal
plan/configuration
Execute system
plan/configuration
and update DT
with field data
1
2
3
4
5 6
7
8
RUN-TIME
DESIGN-TIME
CONTINUOUS
MONITORING
Shut-down or
switchover
CPS
operational
S1 - nominal
no fault/anomaly
S4 - compromised
critical fault(s)
Critical
fault/anomaly
S3 - heavily degraded
multiple faults/anomalies
Critical fault/anomaly resolved
by excluding components
S2 - degraded
minor fault/anomaly
Minor
fault/anomaly
Dependability target
achievable
with current
plan/configuration
Dependability target
non achievable
Dependability target
non achievable
Minor fault/anomaly
Dependability target achievable
with current plan/configuration
Dependability target
non achievable
Critical fault/anomaly
Critical fault/anomaly
recovered
Additional fault/anomaly resolved
Fault/anomaly
resolved
Critical
fault/anomaly
Continuous monitoring and planning and
reconfiguration through DT run-time models.
State-chart describing the transitions among nominal, degraded and compromised
states in self-healing CPS.

20. Francesco Flammini
Towards a hierarchy of Digital Twins
Conceptual architecture integrating the MAPE-K self-healing
loop into Digital Twins at multiple levels of CPS.
Flammini, F.: Digital Twins as Run-Time
Predictive Models for Resilience, Self-
Healing and Trustworthy Autonomy of
Cyber-Physical Systems: a Conceptual
Framework. Philosophical Transactions of
the Royal Society A. Special Issue on
Symbiotic Autonomous Systems. To appear.

21. Francesco Flammini
A real-world example
Francesco Flammini, Stefano Marrone, Roberto Nardone, Mauro Caporuscio, Mirko
D’Angelo, Safety integrity through self-adaptation for multi-sensor event detection:
Methodology and case-study, Future Generation Computer Systems, Volume 112,
2020, Pages 965-981, ISSN 0167-739X, https://doi.org/10.1016/j.future.2020.06.036.

22. Francesco Flammini
Safe dynamic reconfiguration

23. Francesco Flammini
Custom MAPE-K for multi-sensor event detection
Exclusion mechanism: at every step k, the raw data di(k) captured by each sensor i, and the decision
D(k) taken from Voting are used to analyze the system and compute the new sensor reputation rpi(k).
DYNAMIC BAYESIAN NETWOKS
(supporting XAI)

24. Francesco Flammini
Safety integrity levels depending on reconfiguration
SIL requirements for continuous operation in IEC 61508
Francesco Flammini, Stefano Marrone, Roberto Nardone, Mauro Caporuscio, Mirko
D’Angelo, Safety integrity through self-adaptation for multi-sensor event detection:
Methodology and case-study, Future Generation Computer Systems, Volume 112,
2020, Pages 965-981, ISSN 0167-739X, https://doi.org/10.1016/j.future.2020.06.036.

25. Francesco Flammini
Process mining
Caporuscio M, Flammini F, Khakpour N, Singh P, Thornadtsson J (2019). Smart-Troubleshooting
Connected Devices: Concept, Challenges and Opportunities. Journal of Future Generation of Computer
Systems (Elsevier), vol. 111, pp. 681-697, October 2020, doi: 10.1016/j.future.2019.09.004
Singh PJ, Flammini F, Caporuscio M, Saman Azari M, Thornadtsson J (2020). Towards Self-Healing in the Internet of
Things by Log Analytics and Process Mining. In Proc. ESREL2020 - PSAM15, 30th European Safety and Reliability
Conference and 15th Probabilistic Safety Assessment and Management Conference, Venice, Italy, June 21-26 2020

26. Francesco Flammini
Roadmaps for AI Integration in the Rail Sector (RAILS)
https://rails-project.eu/about/ Flammini F, Vittorini V, Zhiyuan L (2020). Roadmaps for AI Integration in the Rail Sector – RAILS. In: ERCIM News, vol. 121, p. 34-35

27. Francesco Flammini
Thank you for your kind attention!
Questions?
“There is nothing permanent except change”
Heraclitus