Beyond autonomy: how AI is driving the automotive industry

CEO and founder of Monolith AI

Artificial intelligence leads many of the most innovative areas of the automotive industry at this time. Although traditionally associated with autonomous cars, the industry is also working on AI applications to revolutionise other areas of the sector including, mobility services, supply chain, predictive maintenance and customer experience, to name a few.

As today’s vehicle becomes more and more complicated to design and build, AI is working behind the scenes to provide a productivity boost, streamlining design and manufacturing elements associated with the sector.

Manufacturing: steady-state track testing and transient testing

Each development phase within manufacturing needs a different assessment method to evaluate a car’s performance: CFD simulations before any prototype is produced, wind tunnel experiments, and eventually, track testing. Generally speaking, track testing provides the most accurate and relevant performance results from which to start the design process. These initial tests are used to direct development towards designs that will have the best performance in steady-state track testing.

Steady-state track testing

Track testing provides large datasets from sensors measuring acceleration, velocity, temperature, and more. Understanding the underlying performance of the car in this “dynamic” environment is extremely difficult. The complex interaction of all the vehicle components, weather conditions, track conditions, and driving style, makes for a huge set of complex high-dimensional data. Instead of testing the dynamic response of the car, some manufacturers test in a “steady-state” manner – for example, driving in a straight line at constant speeds. This makes it easier to extract the car performance from the sensor data.

AI and its role in keeping engineering moving during lockdown

Steady-state testing has its limitations however. For example, most of these tests performed on real tracks are to assess other aspects of the car’s performance (such as the cooling system or the comfort of the driver). While at the track, separate tests are performed to gather data for the steady-state response of the car. Even after this, the resulting performances measured by these steady-state tests are often too simplistic to model the car’s full dynamic response.

The advantages of transient testing

A more elegant solution is to use AI algorithms to learn the performance of the car in “transient tests”. These dynamic-learning algorithms learn the physics of the car by using all the available data from the normal real track tests, without the need for separate steady-state tests. The trained models can be used to predict the dynamic behaviour of the car for unseen cases, such as on another track.

Dynamic models can predict complex behaviours that steady-state models cannot answer. For instance, whereas steady-state models can only answer questions such as “what is my down force at 60mph”, a dynamic-learning model can answer: “how is the down force affected if I accelerate from 30mph to 40mph while taking a sharp left corner?”. Moreover, this test can be combined with other tests that are required by legislation, saving money and weeks of testing time.

Prediction-based AI testing at Kistler 

Kistler, a global leader in dynamic measurement technology for pressure, force, torque, and acceleration applications, has been using AI technology built by Monolith to run track tests to predict the force applied on an automobile's wheels under different circumstances.

Dynamic models can predict complex behaviours that steady-state models cannot answer

A Machine Learning model is trained on existing track test data, which enables them to predict a car's dynamic behaviour for other, unseen circumstances. During the evaluation phase, their model can identify regions of higher uncertainty and communicate these clearly to the user. Using digital testing like this reduces the number of in-person testing days required to accurately assess vehicle behaviour by up to 70% - from 11 days on the track to three.

These prediction-based AI tests have been shown to save engineers' up to 40% of their time that they would have spent on repetitive tasks, whilst also saving on R&D money. Given the increasing amounts of time and resources spent on understanding the condition and performance of remote assets, the prospect that machine learning can be realised using modelling and simulation is an exciting indication of AI’s potential.

AI is driving innovation from behind the scenes

Just about every part of the automotive sector is going through some form of digital transformation, with AI drastically reducing the time for which these innovations are happening.

Whilst the goals will differ depending on what is being achieved, increasing the proportion of development work that occurs digitally rather than physically is something we are seeing across the sector.

So next time you read an article about how AI is fuelling the technological developments in autonomous cars, bear in mind all the unseen ways in which machine learning technologies are advancing the automotive sector behind the scenes.

Dr Richard Ahlfeld, CEO and founder of Monolith AI