Digital Twins for Autonomous Intelligent Systems: From Development to Deployment

TY - GEN

T1 - Digital Twins for Autonomous Intelligent Systems:

T2 - 2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C)

AU - Esterle, Lukas

AU - Ejersbo, Henrik

AU - Frasheri, Mirgita

AU - Gomes, Cláudio

AU - Macedo, Hugo Daniel

AU - Larsen, Peter Gorm

PY - 2022/11

Y1 - 2022/11

N2 - Autonomous systems are operating without human intervention. This requires them to have a good model of themselves and their environment to ensure they operate safely towards their goals. Within a confined or static environment, a simple model can be enough to achieve this requirement. However, if we intend to have autonomous systems to operate intelligently within as well as together with their environment, a model alone will not suffice. Limitations and assumptions of the models and the initial prototypes that informed the modelling process, as well as environments never encountered while in prototyping stage, eventually render the model of a deployed system inaccurate. We can overcome this issue by updating the model at runtime. This is considered a digital shadow as it will reflect the current state of the environment at runtime. When this digital model is not only updated at runtime and reflects the current state but also is able to provide feedback to the real system at runtime, we consider this a Digital Twin (DT). Over the past years, DTs have become not only an essential tool in the development of cyber-physical systems but also an integral part of intelligent autonomous systems.

AB - Autonomous systems are operating without human intervention. This requires them to have a good model of themselves and their environment to ensure they operate safely towards their goals. Within a confined or static environment, a simple model can be enough to achieve this requirement. However, if we intend to have autonomous systems to operate intelligently within as well as together with their environment, a model alone will not suffice. Limitations and assumptions of the models and the initial prototypes that informed the modelling process, as well as environments never encountered while in prototyping stage, eventually render the model of a deployed system inaccurate. We can overcome this issue by updating the model at runtime. This is considered a digital shadow as it will reflect the current state of the environment at runtime. When this digital model is not only updated at runtime and reflects the current state but also is able to provide feedback to the real system at runtime, we consider this a Digital Twin (DT). Over the past years, DTs have become not only an essential tool in the development of cyber-physical systems but also an integral part of intelligent autonomous systems.

KW - Digital Twin

KW - autonomous systems

UR - http://www.scopus.com/inward/record.url?scp=85143075765&partnerID=8YFLogxK

U2 - 10.1109/ACSOSC56246.2022.00028

DO - 10.1109/ACSOSC56246.2022.00028

M3 - Article in proceedings

SN - 978-1-6654-5143-7

SP - 53

EP - 54

BT - Proceedings - 2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion, ACSOS-C 2022

PB - IEEE

Y2 - 19 September 2022 through 23 September 2022

ER -