Achieving sustainability through machine learning
In my formative years I had a thing for sports cars. I was drawn to both classic and modern designs, with my first two cars being an avocado green 1968 Ford Mustang and a fully -loaded silver 1983 Datsun 280zx. Two completely different driving experiences with one common denominator - power.
After graduating from college, my priorities shifted, and I ended up trading in the 280zx for a more practical option. From that moment on, driving became a means to an end - something I needed to do, not something I wanted to do.
Fast forward to today, almost two years after joining Secondmind and relocating my family from the United States to England, cars are once again a priority for me, but for different reasons. When we moved to Cambridge, we bought electric bikes and had every intention of not buying a car. We also had big plans to travel around Europe, but Covid put a stop to all of that within a few months of our arrival. Out of necessity, we purchased an English classic, a Mini Cooper. We weathered the lockdown, and at the first opportunity we hit the road to explore England from coast to coast and many points in between in our trusty Mini. I’d forgotten what it was like to really enjoy driving, and had lost an appreciation for the freedom to explore that my car makes possible. This freedom, however, comes with an added personal responsibility and is a common denominator in other aspects of my life - minimizing my impact on the environment.
I’m not alone. I'd be hard-pressed to find many across my personal and professional networks who either enjoy driving or depend on a car as their primary source of transportation, and who aren’t concerned about their personal carbon footprint. And as consumer demands have increased, so too has the challenge for automotive manufacturers, all of which are grappling with mounting engineering design and production complexities in the midst of a multi-decade transition to zero emissions mobility.
Tighter emissions standards, inefficient production processes and expanding customer expectations have created a perfect storm for the automotive industry, and powertrain engineering in particular. The sheer number of powertrain configurations, including internal combustion engines, myriad hybrids, and pure battery-electric, present an almost insurmountable production challenge in the short and long term. A step change is needed - one fueled by practical, state-of-the-art machine learning adept at solving the most complex optimization problems to dramatically reduce production time and resource consumption.
Despite the apparent flood of battery electric vehicles coming to market, more than 90% of consumers around the globe still plan to buy a petrol, diesel or hybrid electric car as their next vehicle purchase.* And it will be more than a decade before pure battery-electric vehicles are dominant on the road. Driving distance, lack of infrastructure, and high cost remain challenges for consumers, so even those with the best intentions still need that classic Mini to get around.
So, what can we do to help in this transition period to a greener future? Over the past year and a half, Secondmind has collaborated with car makers like Mazda to help them manage the increasing complexity of engineering design to reduce powertrain ECU calibration time. We’re also exploring new, more efficient ways to optimize the development of next-generation powertrains and design processes that will ensure a sustainable future for the automotive industry at large.
My team and I understand the challenges the automotive ecosystem is facing in the long transition to electrification and the complexity of balancing the expectations of customers, the pressures of regulation and the needs of the planet. While there are far more green transportation options now than when I was driving the Mustang, cars remain a crucial mode of mobility for many, and it’s our mission at Secondmind to help the innovators in automotive design cleaner cars, get them to market faster, and to achieve their sustainability goals through machine learning.
Source: Deloitte, 2021 Global Automotive Consumer Study