AI and its role in keeping engineering moving during lockdown

Businesses globally have started realising many operations do not need people in an office as long as the right technology is in place to keep operations running smoothly. Keeping networks secure while employees use remote devices and home Wi-Fi, continuing to collaborate on tasks and managing progress are among the challenges that have emerged while remote working, but AI has been able to aid effective collaboration.

Opinion: Engineer the new normal with AI and automation

Comment: build resilience to beat disruption

Historically, engineers were reluctant to adopt remote work due to a preference for keeping their data local, either for security reasons or speed, but often a bit of both. However, the COVID-19 pandemic has changed many things. For engineers that deal with product testing, international lockdowns have reduced their ability to perform in-person testing, if not stop it altogether.

This is where AI has seen widespread uptake and greater general acceptance amongst engineering professionals. Location-based industries in particular, such as car manufacturers, aerospace and packaging manufacturers, have been harnessing the power of AI to keep their engineers safe whilst keeping their industries moving forward.

For industries normally reliant on in-person testing and simulation, AI brought advantages to engineers and designers in terms of virtual testing capabilities.

Reducing product test cycles

Engineers working on product testing have to design scenarios before most experimentation. Some are crucial safety scenarios, while others are purely about performance. Take the example of an engineer working on a new prototype car model. They would have anywhere between 10 to 100 scenarios to test and collect data on.

• What is the tire traction of my car when it is negotiating a slalom?
• How long does it take for my car to reach 60 mph from a standstill?
• Evaluation of the engine's performance under different scenarios.
• What is the pressure on different parts of the car chassis when taking a sharp turn?

Creating these scenarios helps make the final data easily digestible to answer all the relevant questions. However, this requires lots of time at a racetrack to test and measure—something which is not possible for every engineer during the pandemic lockdowns.

This is where AI comes in. With the right AI platform, engineers can either take data from previous track tests, or with a handful of track test data, engineers can now ask the AI the scenario questions, and it will tell them the most likely outcomes. Engineers at sensor manufacturer Kistler have been increasingly doing this since lockdowns started affecting product tests.

Kistler engineers needed to run a track test to predict the force applied on a car's wheels under different circumstances. A Machine Learning model was trained on their existing tests, which was also able to use this data to predict the car's dynamic behaviour for other unseen circumstances. During the evaluation phase, the model was able to identify regions of higher uncertainty and communicate these clearly to the user. Using digital testing like this reduced the number of in-person testing by up to 70 per cent - from 11 days on the track to 3 - to accurately assessing vehicle behaviour.

Less repetition

By using cases like the one mentioned above, AI helps businesses save R&D money by reducing the number of physical tests by 25 per cent and saving engineers' up to 40 per cent of their time that would have been spent on repetitive tasks.

Today AI solutions are being created and shared at a rapid rate, creating sophisticated systems which link different departments and applications, and enable for more effective collaboration. Throughout this time of separation and working from home, these tools have significantly improved the ability to collaborate on large scale projects. For example, a designer in Silicon Valley and a manufacturer in Thailand can use an AI platform to share expertise, collaborate in real-time and manage vast sets of complex data.

It is a fact that if the pandemic happened years ago and there was a need to conduct physical testing on a car to see how it would behave on a certain track, we simply would not have been able to do so. As the engineering industry starts to embrace new capabilities allowed to us due to AI, we hope to come out of international lockdowns with more sophisticated cloud systems in place, along with a greater awareness and knowledge of how AI tools can help. Any tools which help companies collaborate, share knowledge and innovation can only be a good thing, helping to push the industry into a bright and more connected future.

Dr. Richard Ahlfeld, CEO and founder of Monolith AI