AWS Quantum Computing Blog

Exploring industrial use cases in the BMW Group Quantum Computing Challenge

Today, the BMW Group launched a global open innovation challenge focused on discovering potential quantum computing solutions for real-world use cases: The BMW Group Quantum Computing Challenge. We are delighted to collaborate with BMW on this challenge, and to invite the quantum community explore new approaches to industrial applications. It’s still early days in quantum computing, and the BMW and AWS teams share a goal to enable research and discovery in this nascent field.

“Just as we are accelerating data-driven innovation with AWS across our organisation, we are looking to fuel innovation for the future in quantum computing with the help of the broader community,” said Peter Lehnert, Vice President BMW Group Research, New Technologies. “Different research and commercial organizations around the globe are breaking ground in quantum algorithms, software, and hardware. We aim to tap into that additional innovative power, inspire new thinking, and create opportunities for quantum builders to work with BMW on meaningful business problems.”

Source: BMW Group

The BMW Group Quantum Computing Challenge is open to participants from research groups and companies worldwide. You can register now, and prepare to submit your entries before the September 24 deadline. The challenge is organized into two rounds. In the first round, participants need to submit a well-documented concept proposal for one of four use case challenges, described below. In the second and final round, teams with the top three submissions in each use case will be asked to build out their solutions. The final, virtual presentation to the competition’s judging panel, including domain experts from BMW and AWS will take place in December. The winners will be announced at the Q2B quantum computing industry conference (Dec. 7-9).

To develop and test submissions, participants will have access to Amazon Braket, the AWS fully managed quantum computing service. Amazon Braket provides a development environment to explore and build quantum algorithms, test them on quantum circuit simulators, and run them on different quantum computers, including devices from D-Wave, IonQ, and Rigetti. AWS is offering credits for the use of Amazon Braket and related services to registered teams, upon request.

BMW Group Quantum Computing Challenge Use Cases

The BMW Group Research, Technology, and Innovation team engaged the Amazon Quantum Solutions Lab Professional Services team to help design and support the competition. We worked closely with BMW to define appropriate use cases for the competition, from the many computational challenges that exist inside complex automotive engineering, manufacturing, and logistics domains. The BMW-AWS team had a few key considerations for selecting the use cases. These included the feasibility of using a quantum or nature-inspired approach, the delta between existing classical computing solutions and potential future quantum-based approaches, and the practicality for a time-limited competition format. Based on these, the team developed the following four challenges that have the potential to drive real-world innovation for BMW.

Use Case: Pre-Production Vehicle Configuration

Before new vehicle models can be released for official production at scale, a number of tests have to be performed on a certain number of vehicles that must first be assembled. Required tests include the evaluation of the ability to manufacture components, known as “producibility,” as well as functionality checks of the vehicles. All of this yields a list of requirements, which specify what components need to be produced, how often, and in what combinations. To reduce costs, these requirements should be fulfilled with the lowest possible number of vehicles. This can be achieved by optimizing the combinations of special equipment to fulfill the most possible test requirements within a single vehicle.

BMW robotic assembly line

Source: BMW Group

The challenge for competition participants is to define a way to optimize the configuration of features for a number of cars, such that all of the constraints and requirements can be met — using nature-inspired optimization approaches and quantum algorithms that can run on today’s or tomorrow’s quantum computers.

Use Case: Material Deformation in Production

During the design of vehicle components, each component needs to be evaluated for producibility. One reason this is important is to avoid costly problems that could arise during actual production. For example, a new part design may not be able to be implemented properly by existing tooling, or materials may degrade as part of the assembly process. To avoid such issues, manufacturers use virtual simulation tools to reduce costly tests of physical hardware. BMW uses numerical simulations of material deformation in the preproduction phase of vehicle component manufacturing — for instance using simulations based on Finite Element Methods to predict material properties during the manufacturing process. This use case challenge is to develop a novel quantum algorithmic approach or perform a novel analysis of an existing algorithm to model mechanical behavior.

Use Case: Vehicle Sensor Placement

Modern vehicles come with many sensors, to help provide safety and convenience to drivers. The number of sensors per car is expected to increase even more as autonomous driving becomes more common. Vehicles need their sensors to gather data from as large a portion of their surroundings as possible, but each additional sensor adds additional costs, so the goal is to optimize the positions of sensors to allow for maximum coverage while keeping costs down. Current approaches to vehicle sensor placement use genetic algorithms, for example. The challenge is to find an optimal configuration of sensors for a given vehicle such that the vehicle can reliably detect obstacles in different driving scenarios – using quantum computing or nature-inspired optimization approaches.

Cars with sensors in BMW's Autonomous Driving Campus

Source: BMW Group

Use Case: Machine Learning for Automated Quality Assessment

In the past few decades, quality control has drastically shifted from manual examination of vehicles towards automated inspection. Specifically, machine learning techniques have revolutionized the quality control process, with convolutional neural networks (CNNs) setting new standards in image processing for automated quality inspection. For example, today, manufacturers use machine learning algorithms heavily to assess vehicle parts for cracks and scratches caused by the metal-forming process, based on the successful segmentation and classification of imperfections. Of course, any technology has its limitation, and for CNNs, it is arguably the computational power consumption. As high-performance CNNs usually digest very large datasets, data centers ultimately end up with large and expensive GPU workloads. The challenge for the competition participants is to explore novel quantum, nature-inspired, or hybrid classical-quantum machine learning (QML) approaches, that have the potential to provide faster and more efficient training with higher accuracy, to help improve the automated assessment of vehicles’ quality. Competition participants can explore this use case using Amazon Braket’s integration with PennyLane.

How to get started

To learn more and to register for the challenge, visit the BMW Group’s Open Innovation website. Each use case is distinct and requires unique approaches, so we provide more detailed information on submission requirements for each of them. If you take up the BMW Group Quantum Computing Challenge, don’t forget to request AWS Credits, so you can try out your approaches. We look forward to your submissions!