Compared to conventional computers, quantum computers are able to solve problems faster and more efficiently. Doesn’t this make them predestined for use for functionally safe products? This article outlines both disciplines, highlights the main differences, and provides an outlook on potential applications of quantum computers for functionally safe products.
Requirements for functionally safe products
A wide range of risk-reducing measures in functionally safe products ensures that the residual risk of danger to life and limb is brought down to an acceptable minimum. These measures include requirements for the development process, safeguards against failures and defects in the hardware used, and requirements for the product in terms of determinism or reduced complexity.
Quantum computer specifics
Compared to conventional computers, quantum computers are characterized by the fact that they don’t just calculate with the bit states 0 and 1. Rather, they are able to compute simultaneously with all intermediate states (cf. superposition, Schrödinger’s cat, and wave–particle duality). Due to these properties, quantum algorithms are part of the class of probabilistic algorithms, i.e., one obtains results with a certain probability. To obtain statistically reliable results, quantum algorithms must be run multiple times. The susceptibility to errors (noise) makes it necessary to protect against errors in the intermediate steps of the algorithms. The decoherence time has a great influence on the size and execution time of quantum algorithms – to ensure the end result isn’t just noise.
Discussion of possible uses of quantum computers for functionally safe products
A major aspect of developing functionally safe products is determinism: at any time, it must be possible to clearly identify how the system is behaving. This is in direct contradiction with the use of probabilistic algorithms, which yield results only with a certain statistical probability. In principle, this does not exclude incorrect conditions or conditions involving danger to life and limb. The concept of superposition and thus the application of superimposed states further supports this statement.
Another main aspect of functionally safe products lies in securing the hardware via software. In quantum computers, we are currently in the NISQ (noisy intermediate-scale quantum) computer era, meaning that currently available systems are characterized by short decoherence times and high likelihood of error. Consequently, current quantum algorithms may only require very short runtimes on these systems to get results that are not just noise. Furthermore, only about 10 percent of the available qubits (the smallest computational and information unit of a quantum computer) can actually be deployed for calculations; the rest is needed for error checking.
Due to the increasing networking of systems and subsystems, in the future cybersecure communication will be of crucial importance for functionally safe products. Today, protocols for quantum key exchange that use the principles of quantum computing are already available. Following a call for proposals, the U.S. federal agency National Institute for Standards and Technology (NIST) last summer selected the first postquantum encryption algorithms that address the problem of calculating the private key in asynchronous encryption and are thus supposed to be safe against quantum computers. This is made simpler mainly through Peter Shor’s algorithm for quantum computing.
AI (artificial intelligence) is also finding its way into functionally safe products. Various standardization committees are currently addressing the question of what requirements should be placed on AI in this context. Quantum computers are hoped to provide shorter training times or better results for AI. Only time will show whether and to what extent the expectations in this broad field of research will be realized.
Conclusion
Considering all these facts, quantum computers are only likely to be applicable for the development of functionally safe products in the (distant) future.
Offline, they will potentially be used for faster AI training, for example for ADAS (advanced driver assistance systems) in the automotive sector, where both environment recognition and path planning/navigation can be supported. It is also conceivable that they will be used in production planning or for multicriteria optimization/vector optimization for products. Results could then be obtained more quickly and more criteria could be considered than with conventional computers.
Last but not least, online systems are also conceivable as built-in components of a product, e.g., for secure communication via quantum key exchange.
To be prepared for the future, infoteam is looking closely at the quantum technologies currently available and their application options. This consideration is taking place in intensive exchange with experts from other application fields, such as functional security and cybersecurity.
author: Marc Maußner, Senior Engineer, infoteam Software Gruppe
https://infoteam.de/karriere-bei-infoteam https://infoteam.de/unsere-maerkte
https://infoteam.de/pressebereich
foto: AdobeStock Foto DB
