Autocar Drivers of Change 2020: winners revealed

“The harmonics produced from charging EVs in the UK are currently not significant,” he says, “but what will happen in 20 years when the number of EVs massively increases?” The harmonics could be “a real issue”. This is a problem because harmonics lower power quality, in turn causing charging problems and damaging expensive transformers and other equipment. The cost to the UK alone could be millions of pounds, Costa Capezzone says.

So what’s the solution? Costa Capezzone’s idea lies in smart cities, smart grids and his invention, Smart Measurement Points. Taking as a base a device he developed in collaboration with the University of Valladolid, which measured the harmonics produced by EVs such as Nissan’s Leaf, these would comprise a network of artificially intelligent packs, installed in homes, that would measure how an EV is charging. By communicating with each other over a 5G network, they would then correlate this with the status of the grid and automatically adjust EV charging times and other consumption to prevent overload and minimise harmonics.

Costa Capezzone impressed the judges with his “great depth of knowledge”, “passion” for the topic and “forward thinking”. His next step is to test his technology in a medium-sized city such as Oxford, ideally in partnership with a local autonomous car firm.

Finalists:

Modular in-built wireless charging points in assigned parking spaces for electric vehicles - Kelsie Osborne

EV turbine charging, increasing a vehicle’s range in an environmentally friendly way - Gurjit Sandhu

Digital winner

Automated offensive and defensive security (automods) platform for cars and self-driving cars - Dr Madeline Cheah

The winner of the Digital category is Horiba MIRA’s cybersecurity R&D lead, Dr Madeline Cheah, with Automated Offensive and Defensive Security (Automods), a platform to protect cars, including self-driving cars, from cyber attacks.

While innovations like increased connectivity and autonomous cars make vehicles more user-friendly, Cheah explains, they also make them more vulnerable to hacking, which could be disastrous for self-driving cars. The defences that currently exist are mainly “static and updated only intermittently” and the problem is compounded by a shortage of car cybersecurity experts. There is thus a need for sophisticated defences in the car.

Cheah’s solution has four parts. It begins with ‘attack trees’, blueprints for kinds of potential cyber attacks, initially created by Cheah and her team at Horiba MIRA but eventually automated. These are then fed into a program that uses the trees to simulate attacks and feedback on the results. From this, machine learning can then be used to ‘teach’ the program to identify intrusions automatically and find ways of defending against the attacks.

The concept has its technical roots in Cheah’s PhD thesis on automotive cybersecurity but, innovatively, she stresses that some of its principles are “grounded in biology”. The final stage will comprise an ‘artificial immune system’ with two layers. ‘Innate immunity’ will allow the system to identify intrusions and destroy them. However, because this technology is “very heavy-handed” (to counteract battery attack, it might “cut power to the vehicle”, say), Cheah proposes adding another layer of ‘adaptive immunity’ to allow its responses to be “more gradual and proportionate”.

The judges praised Cheah’s presenting skills and her innovation in identifying a “very large, complex real-world problem” for the industry and potentially solving it – feedback that, Cheah says, has given her the confidence to take her idea forward commercially. Cheah now plans to begin commercialising the first phase of the Automods platform, releasing it as a subscription-based service to selected early adopters, from midway through next year.

Finalists:

A smart meter for your car that measures its environmental impact, including emissions from manufacturing - Tom Harle

An open mobility platform that merges car buying, admin and selling in an omnichannel app - Luke Greengrass

Retail winner

Carbusi car model recognition and car buying app - Marvin Samuels

Many customers struggle when buying a new car. They are daunted by technical details and have trouble cutting through jargon. This means they often find car buying confusing and have difficulty choosing the right model, especially when there isn’t an expert or enthusiast on hand to answer their questions, says the Retail category winner, Marvin Samuels.

Samuels’ idea, which draws on his background as an IT consultant and a car enthusiast, is a new car buying app to solve this problem. Carbusi (from ‘Car Buying Simplified’) is modelled on the popular music app Shazam, but instead of identifying a song from its sound, it identifies cars from a photo and then guides users through the process of purchasing it, simplifying car buying so that even people with little knowledge of cars can understand.

After identifying the car, the app will pull up details of the model, including specifications, equipment and price range. It will also highlight the closest dealerships and car sellers. The app will give information about ongoing costs such as road tax – “often overlooked” by buyers, says Samuels. Finally, an augmented reality feature will allow users to see how the cars look in their driveway. Samuels says the app will work for new and used cars, so long as he can obtain their data for an image and car database.