The need for availability has resulted in designing robust redundant systems in automotive domain. Introducing hardware redundancies is a traditional approach to achieve fail-operational systems which increases the cost, power and size of the actual item under control. Achieving redundancy through digital twins is an efficient solution which address the problem of having additional hardware. Virtual models of the system under control are deployed either in Edge or Cloud as part of the digital entity. The models interact continuously with the system and provides various controls including availability in case of a fault. But when the digital twins are deployed in safety automotive control like steering or chassis systems, the challenge of developin g and using it in run time for decision making without any failure is very critical to avoid hazards at vehicle level. This paper explores the challenges and propose a safe architecture while deploying a digital twin in safety use cases.
Krishnamurthy et al. (Wed,) studied this question.