Efficient crash test simulations using quantum surrogate models
Q-Surrogate
Q-Surrogate - Digital twins Integral safety DLR prototypes
Computing-intensive simulations are integral to the development of innovative and safe vehicle concepts. Quantum computers can ensure that these are faster and more precise.
Efficient crash test simulations using quantum surrogate models
High-precision simulations are increasingly replacing complex tests, significantly reducing the effort involved in crash testing. The 'Quantum Surrogate Modelling' (Q-Surrogate) project aims to speed up these calculations and simulations by harnessing the power of quantum computers. The surrogate models developed as part of the project enable fast and efficient approximations to be made to various configurations of the test setup.
One of the biggest challenges in modern simulation projects is ensuring transferability between different test configurations. In vehicle crash tests, for instance, various materials, geometric parameters and environmental conditions must be considered. At Q-Surrogate, we are therefore developing AI-based surrogate modelling methods to quickly cover the parameter space. This allows suitable configurations for real experiments to be assessed quickly and efficiently. In addition to crash tests, we are planning use cases in the field of cyber security. Our approach is based on Gaussian process regression, a method that has already been proven effective in various applications. However, this method is not scalable with large data sets. This is where our quantum algorithm comes in, overcoming this bottleneck. An initial quantum algorithm was developed for this purpose in the ELEVATE project. Our aim is to further develop this approach in Q-Surrogate to prepare it for application-relevant datasets.
Our methodology has several advantages over other quantum methods:
smaller quantum circuits
no need for a fully error-corrected quantum computer.
The aim of our project is to facilitate progress in a variety of industrially relevant applications, especially in crash testing. We are also employing novel methods to draw conclusions about the application of quantum AI in cryptography. The results will be tested in use cases and the scalability to economically relevant applications will be assessed through the planned integration into existing software frameworks during the project period.
Contribution Institute for AI Safety and Security
As the leading institute, the Institute for AI Safety and Security is particularly responsible for further developing the method designed in ELEVATE. We are combining this with findings from our Quant²AI and NeMoQC projects. We are evaluating new findings in cyber security. Specifically, we are investigating the potential for quantum AI-based replacement models for the cryptographically relevant Shor algorithm.
You can find our current paper on this topic HERE.
