By Jonah Sachs
Year after year, quantum technology continues to boom. In the process, more industries are venturing into the quantum sector, one of which is the automotive industry. Companies such as BMW, Mercedes, and Volkswagen have partnered with quantum hardware and software companies to determine the possible applications for quantum technologies to the automotive field. In the last few years, they have produced marginally better results than the classical computational alternatives. In the long term, they envision an industry highly impacted by quantum technologies.
NISQ vs FTQC: Different Eras of Computation
No matter their modality, current qubit models are very noisy compared to what is required for sophisticated computation. This prevents current quantum computers from being used for general computation. As such, this era of quantum computers has been dubbed that of Noisy Intermediate Scale Quantum Computing (NISQ). Quantum manufacturers employ both hardware and software techniques to amend these errors. A hardware-based example comes from the French company Alice and Bob, who can physically prevent bit flips at an exponential rate with their cat qubit design. For algorithmic error suppression and correction, Google just recently notably announced surface code below the surface code threshold, proving the scalability of these types of designs. These techniques aim to produce machines large enough and error-resistant enough to provide general computation for areas with possible quantum ad- vantage over classical systems. This type of system, full of error-resistant (or logical) qubits, will be that of Fault Tolerant Quantum Computing (FTQC). Companies have varying estimates for how long it will take to reach the FTQC threshold.
Numerous applications for quantum systems have been proposed, some with computational requirements more sophisticated than others. These ideas offer theoretical advantage over classical techniques due to the exponential scaling of quantum information, but require different qubit counts, circuit depth, or levels of error-resistivity of quantum systems to maintain advantage over classical approaches. The main ones, specifically with automotive applications, include optimization problems, numerical simulation, quantum chemistry, quantum ML, quantum cryptography, and quantum sensing.
Near Term Applications and Current Collaborations
Many major automotive players have rushed towards partnerships with quantum computing companies. These partnerships have centered around a few core technologies which are applicable on NISQ systems. Typically, these problems do not require precision calculation, and often employ many shots, or repeats, of the same algorithm to produce a desired result.
One of the most popular application involves optimization, the maximization or minimization of a complex system of variables. Many optimization problems do not possess polynomial-time solutions, and there are examples of faster results achieved using hybrid quantum-classical algorithms, compared to the classical alternative. These methods are referred to as variational, because they typically use a classical optimizer to change parameters within repeated shots of a quantum gate or set of circuits. Often, these algorithms can be used as a step in processes for computational chemistry, solving for the energy of a system, or to solve optimization problems in fields such as logistics and finance.
Automotive companies have rushed to use these types of algorithms. A few years ago, Ford and Quantinuum partnered to use fewer resources for variational algorithms applied to the molecular modeling of lithium ion batteries. Quantinuum also worked with BMW and Airbus to model fuel cell catalyst re- actions, also for battery development. Similar to BMW, ion-trap company IonQ partnered with Hyundai to use variational algorithms for catalytic simulations in the development process for batteries. Mercedes has pursued something similar, partnering with IBMQ and Google Quantum AI for battery development using variational algorithms. More recently, Volkswagen and superconducting qubit company IQM announced a new algorithm improving the accuracy and scale of variational algorithms for chemical modeling. There is immense interest in this field, with further applications in the FTQC era, as will be discussed later. Richard Feynman dreamed of modeling atomic systems with atoms, car companies are trying to pursue this same dream, developing the atoms of batteries with atoms.
Automotive companies are also using variational algorithms for other optimization use cases, besides chemical modeling. BMW partnered with AWS to work towards optimizing robot trajectory planning within factories. This is an example of solving the traveling salesperson problem: a famous NP-Complete example of optimization with non-polynomial optimal algorithms for true instances. BMW is not alone in this endeavor; many large corporations are using variational algorithms and quantum approaches to optimization to deal with logistics problems. Solutions to these types of logistics issues can often be extremely cost-beneficial, UPS is reported to have saved between $300-400 million dollars annually from making the decision to (almost) never make left turns. This was not a problem solved via quantum mechanics, but simply illustrates the impact solutions to optimization problems can have on profit. Another company pursuing this type of logistical optimization is Volkswagen. They partnered with D-Wave, a company specializing in optimization with quantum annealers, to produce bus routes which take less time, and use less fuel by avoiding city congestion. D-Wave has also worked with Toyota towards a similar traffic-based problem, this one instead involving a 4-way traffic light intersection. Optimization has also been used for vehicle architecture and routing, notably Classiq, NVIDIA, and BMW.
It seems that variational and other hybrid algorithms will be a major factor in the near-term quantum automotive industry, both in computational chemistry and other complicated data analysis applications, including logistics. Results have been achieved past the classical barrier, but only for specific problem types. These processes seem to be stepping stones towards dreams of atomic and materials modeling without simplification, a goal only possible with extremely large FTQC systems, and never imagined on classical computers.
In addition to these variational methods, there are other areas of overlap between the automotive industry and quantum technologies which are notably making progress as well. One area of interest to automotive companies is numerical simulation. Numerical simulation for automotive vehicles, including those for computational fluid dynamics (CFD), can often take extremely long periods of time. There is hope that a combination of classical and quantum systems could be used to increase the speed and minimize the resource allocation of these procedures. An example is BMW, who has partnered with neutral-atom company Pasqal to work towards improvements in partial differential equation modeling. Specifically, they published results surrounding numerical simulations of the modeling of metal forming. IonQ also recently partnered with engineering design company Ansys in a similar vein, aiming to bring quantum advantage to the field of engineering simulation.
There have also been some developments made in the overlap between the AI/ML and quantum technologies. Most notably, IonQ and Hyundai partnered to use Quantum AI for computer vision for self-driving. Many think of quantum AI as more of a long-term application for quantum systems. There are many applications for AI, and thus quantum AI, to automotive systems, and the area has been well studied.
Quantum cryptography is another area of interest for automotive manufacturers, especially as the amount of technology within vehicles increases. Similar to other corporations, automotive companies will need to switch to Post- Quantum Cryptography schemes (PQC) to avoid the threat of Shor’s algorithm on current encryption mechanisms. Quantum communication companies are currently developing quantum secure communication technology, which will be applicable to automotive communications as well. The Cybersecurity and Network Defense Research Agency, among others, has recommended caution around quantum cryptographic attacks, especially as we progress toward autonomous systems.
Finally, there has been a lot of short-term interest between automotive manufacturers and the quantum sensing industry. Through the laws of quantum mechanics, quantum sensing companies have promised more accurate sensors. Examples include magnetometers, gyroscopes, gravitometers, accelerometers, and other sensing devices. Specifically, as an example, a quantum atomic magnetometer, using the alignment of the spin of a groups of atoms, could prove more accurate than classical alternatives, while also protecting against GPS spoofing. Automotive companies buy and utilize tons of sensors, it is clear why there is interest in more precise (and sometimes cheaper) technology. In addition, quantum sensors require less precise hardware than quantum computation systems, and are more of a near-term technology than many quantum computing applications. A proof of concept for this idea comes from Capgemini, who has used quantum sensing to improve fault detection in materials. McKinsey also recently noted that quantum sensing could offer many untapped advantages over traditional sensing systems, specifically in mobility-based industries.
Long Term Aspirations for Quantum Computers and the Automotive Industry
But what about the long-term, in the era of the FTQC? We can only theorize about the possible applications in the automotive industry, but the implications are wide.
Most cited, as briefly mentioned earlier, are the applications for atomic, molecular, and materials modeling. Variational methods can approximate an energy level using repeated measurements of a simple system and some classical resources. Given an FTQC system, no shortcuts will be required, and full molecular/materials modeling could become possible. However, this will require a geometric interpretation imported onto a quantum computer, which is not currently possible. Once this modeling becomes feasible, it could lead to improved materials development, specifically optimized for vehicle production in areas like weight management and for aerodynamic capabilities. There have been recent developments in loading non-uniform datasets, a crucial step towards creating geometric models on quantum systems. There has also been work specifically surrounding the optimization of geometric objects on quantum systems. Quantum numerical simulation has been well studied; the applications are wide across many industries, past the simple differential equation solvers discussed earlier. There is a reason numerous companies are pursuing this quantum application, despite limited advantageous results produced during the NISQ era.
In addition, quantum communication systems will be leveraged in future automotive vehicles for their speed and security. One can imagine a world where the majority of information transmitted within a vehicle, between sensors and computing systems, is quantum. Vehicles with an internal “quantum internet” would possess many advantages compared to the current classical communi- cation systems used within automotive vehicles, in both security and speed. These systems would also be able to transmit more secure protocols with other quantum vehicles, in a proposed quantum-driven internet of vehicles, improving the safety and driving capabilities of an interconnected network of many vehicles.
As we progress into the FTQC era, there will be further developments and improvements in NISQ-applicable areas of discussion with classical analogs. These include optimization, numerical simulation (as discussed above), and quantum AI. As our qubit count, coherence time, and 2-qubit gate error rate improve into the FTQC era, so will our ability to improve these mechanisms into quantum advantage over classical systems. I predict a future where super- computing and quantum computing are used hand in hand to solve these com- plex problems, while also limiting the energy consumption of future hybrid super-computing systems. So does NVIDIA, who has invested heavily in their relationships with companies producing quantum technologies, combining their super-computing GPU architecture with quantum operating systems.
Conclusion
Inevitably, I’m sure to have missed some major players in the automotive and quantum space, and possibly even some applications. Please let me know any that may have escaped my research! The field is ever evolving, and it is impossible to know where transformative technologies will allow us to explore, and all possible applications. I am confident that the automotive industry will be revolutionized by quantum in the long-term. So are many automotive players, as they are funneling resources into NISQ systems despite little evidence of quantum advantage, or, at best, marginal advantage.
Four years ago, the question was whether a fault tolerant quantum system could be created. After Google’s announcement, the question is now when they will be created. Automotive players know the potential power of the FTQC systems, and are investing hard before the technology has reached its maturity, or even much practical advantage. FSAE may be a long way away, but look out F1, quantum is coming.
About the Author
This piece was written by Jonah Sachs, with assistance and encouragement from Doug Finke. Jonah is a senior at Washington University in St. Louis. He is the lead of the Electronics and Data Acquisition team for a Formula Student vehicle (FSAE), and is also the president and treasurer of WashU Club Golf. Jonah is interested in quantum technologies, semiconductor fabrication and physics, and electrical engineering and prototyping. He is currently pursuing research in the fabrication of Josephson Junctions, utilizing multiple lithography mechanisms for the production and calibration of parametric amplifiers. Jonah can be reached at [email protected].
January 2, 2025
Leave A Comment