Smart models for smarter decision making


Consider, for example, automated driving systems. While autonomous vehicles promise to dramatically improve mobility, engineers must test these frameworks for critical factors such as safety and potential system failures. Toyota is one of the automakers who strive to secure driverless systems. In 2016 Toyota President and CEO Akio Toyoda Said More tests would be necessary to complete its mission – about 8.8 billion kilometers.

Fortunately, says Stefan Jockusch, vice president of strategy at Siemens Digital Industries Software, simulation can help. By virtually testing millions of real-world scenarios, from snowy road conditions to reckless pedestrians, simulation technology can analyze the performance of autonomous vehicles while speeding up development and reducing costs.

But while simulation is essential to the digital development and manufacture of products today and tomorrow, challenges such as increased complexity and a lack of domain knowledge are pushing organizations to strengthen their simulation processes with artificial intelligence (AI) capabilities.

AI have smart increase

While the challenges may vary, Don Tolle, director of consulting and research firm CIMdata, says: “One of the biggest obstacles to simulation is the time it takes to transform a complex simulation and share the results with others, including the design. simulation engineers and analysts. In fact, Tolle says that designing, gathering information, creating, running and analyzing what-if models to support decision-making can take “weeks.”

Complexity is another hurdle that engineers face. Simulation models can provide deeper and more precise information about the behavior of manufacturing systems, but these additional details can come at the cost of greater calculations. Building simulation models also requires talents with in-depth knowledge of the field and mathematics. Many organizations are focusing on democratizing access to simulation tools by making them a standard part of design and manufacturing processes. But the challenge, warns Tolle, is “to make these tools usable by the average engineer who may not have extensive field knowledge in the specifics of simulation and simulation technology.” After all, developing AI algorithms is only part of the simulation process; engineers need domain knowledge to understand the larger context of how models are built and the purpose they serve.

In response to these hurdles, many companies are turning to AI to speed up and simplify simulation – and for good reason. AI can distill information into a form that’s easier to understand and more transparent for engineers, eliminating the need to interact with every detail of a model. “The ability to create these incredibly complex models is one of the areas where artificial intelligence and machine learning will have the greatest impact,” says Tolle.

Indeed, AI can “learn” its expertise from the vast volume of simulation datasets created by thousands of simulation runs in similar applications. As a result, AI can come up with model parameters that allow an optimal set of design features for the system while eliminating the risk that simulations take longer than physical tests. Then engineers can begin to put together optimal design features for more detailed designs, such as 3D computer-aided designs, software development, and electronics. “Simulation increases the intelligence of the engineer by using AI and [machine learning] to improve the way we perform analysis and use data, ”says Tolle.

No shortage of use cases

AI can help make simulation practical in cases where it wouldn’t be otherwise, for example, when a designer wants to quickly test and validate many design configurations.

“Simulations can be computationally expensive – for example, the charging behavior of a hybrid electric vehicle for thousands of types of drive cycles,” Jockusch explains. AI helps develop so-called surrogate models, using thousands of existing simulations to derive highly simplified models, much less computationally expensive, that are “precise enough to guide designers through a complex decision space”.

Another benefit of AI is its ability to detect design flaws early on in a product’s lifecycle. “There have been notable examples of system failures or system errors over the past four or five years in the aerospace and automotive industries with recalls and major issues,” says Tolle. “The cost of making decisions late in the life cycle is enormous.”

The good news, he says, is that AI can minimize the risk of introducing loopholes in product design by allowing engineers “to validate systems throughout their development.” This allows for smarter, faster design decisions and tradeoffs early in the design lifecycle rather than having to change the design later, which can be costly in complex systems. ”

Download the full report.

This content was produced by Insights, the personalized content arm of MIT Technology Review. It was not written by the editorial staff of MIT Technology Review.

Leave a Reply

Your email address will not be published. Required fields are marked *