UK-based artificial intelligence (AI) software company Monolith is looking to radically reshape automotive engineering. It has created a solution it says can cut automotive development time and costs by up to 50% using machine learning to predict results more quickly.

Founded in 2016, Monolith is led by CEO Dr. Richard Ahlfield. He and his team, which includes Saravanan Sathyanandha, CTO; Rebecca Geier, CMO; Dr. Joel Henry, Principal Engineer; and Peter Wooldridge, Head of Data Science, believe that the company is set to rapidly scale its technology thanks to £10.6 million in funding from some of the world’s top software investors.

Its software uses self-learning models to quickly predict the results of complex vehicle dynamics systems, reducing the need for physical tests or simulations. The approach is said to dramatically accelerate every stage of the automotive development process—from initial design to design iterations, validation, and production, which the company describes as currently requiring repetitive, time-intensive, and costly tests and simulations. The use of its platform is also said to result in fewer physical prototypes, travel to specialist test sites, and on-road testing, making the latter stages of validation safer and more sustainable.

 

Transforming simulations

Virtual simulations help reduce the number of physical tests required, but the accuracy and fidelity of the results can be limited. Numerous physical tests are therefore still needed to calibrate and validate the virtual results, as well as to understand performance in operating conditions that cannot be simulated.

According to Monolith, virtual and physical tests create significant volumes of valuable data that are presently underutilized. The data can be leveraged to train highly accurate AI self-learning models to quickly predict the performance of systems by understanding their behavior from data instead of solving the complex physics of the system or performing physical tests.

The company’s solution can harness data from across the entire product engineering process. On its website, it lists compatible applications like those from Altair, Ansys, Autodesk, Dassault Systemes, MathWorks, Siemens Digital Industries Software, and SimScale. For instance, Monolith partners with Siemens Digital Industries Software to facilitate tight integration of Simcenter data sources and simulation environments with Monolith’s software platform.

Using the Monolith approach, engineers can rapidly predict performance in more operating conditions and for areas of the car that were previously impossible to simulate, further reducing the amount of testing required. Its solution is already being used to improve wind tunnel, track, wheel/tire, vehicle dynamics, durability, crash, and powertrain testing.

 

Resonating with engineers

For example, optimizing the aerodynamics of a vehicle to reduce drag has been notoriously difficult to solve mathematically, which reduces the accuracy of simulated models. Owing to the highly iterative nature of the automotive design process, engineers supplement virtual aerodynamics testing with hundreds of hours of wind tunnel tests in facilities that can cost thousands per hour.

“Monolith was founded to empower engineers with AI to instantly solve even their most intractable physics problems,” said Ahlfeld. “We know this resonates especially with automotive engineers who struggle to optimize hundreds of often conflicting criteria with hundreds of complex simulations.”

Requiring hours or days to solve, he says that engineers have grown frustrated by the considerable amount of physical testing still required to make up for the limitations of the virtual tests.

“At the same time, the data that is created in the process represents an enormous opportunity when used with AI,” he said. “By predicting results with self-learning models, we can radically accelerate the development process.”

Today, automotive companies are spending billions developing electrical architectures and software capabilities as they strive to win the race for electric, shared, and autonomous mobility. This squeezes R&D budgets and product timelines in other areas, creating enormous pressure on the engineering teams working to develop higher quality vehicle hardware systems in less time and with fewer resources.

“As Akio Toyoda, CEO of Toyota put it, ‘data is the new gold,’ but the ‘[vehicle] platform will be the backbone for mobility as a service, for autonomy, for car sharing, for any number of services that we want to make possible,’” recounted Ahlfeld. “Data to make better vehicles whilst cutting costs and saving time; this is at the heart of how Monolith is uniquely transforming vehicle development.”

 

Maturing the technology

Monolith has spent the past six years working closely with some of the world’s top engineering teams to develop its platform. Now its technology is being integrated into customers’ activities, with engineering teams at leading automotive OEM and Tier 1 suppliers said to be already realizing substantial reductions in physical testing.

It gives a few examples. Sensor and instrument company Kistler achieved a 72% reduction in sensor-based testing. Honda recorded an 83% faster design cycle. JOTA Sport’s endurance racing team reduced the number of simulations and tests by 50% and associated costs by 66%.

“Optimizing a system, or finding a new solution based on a decade of historical data, is like instantly offering an engineer a decade of experience,” said Henry. “That’s the power of AI; it supercharges an individual’s subject matter expertise by unlocking the expertise stored within a company’s data.”

The no-code platform’s use cases are dependent on the needs of the business and the type of data.

For example, an OEM can use its legacy data to find new insights hidden within its decades of expertise and unique data. Alternatively, data captured from a handful of tests using a physical prototype can be used to teach Monolith’s self-learning models to predict behavior over more operating conditions—including under non-steady states when variables of interest are still changing over time.

Monolith self-learning models predict behavior under these typically difficult-to-capture non-steady states in a matter of seconds instead of weeks or months for all driving and operating conditions. This enables engineers to explore even more parameters and requirements to make products that are even more fit for purpose while substantially reducing development time.

 

$46 billion opportunity

Monolith’s business is currently focused on automotive customers, but the company has ambitions and sees applications in many industries. Its solution can be used for any system that requires data, repetitive testing, or digital twins for design development, validation, production, or data evaluation.

Digital twins, real-time virtual representations of a physical object or process, are increasingly used in a range of industries including manufacturing, healthcare, supply chain, and retail. The digital twin market is estimated to be worth over $46 billion by 2026.

Monolith is already working in this space with global brands such as L’Oreal and pharmaceutical company Nanopharm.