Automated capabilities and AI in the vehicle: Status and expectations

14 Mar 2019 01:00h

Event report

[Read more session reports from the Symposium on the Future Networked Car (FNC-2019)]

The session was moderated by Mr Roger Lanctot (Director, Automotive Connected Mobility, Strategy Analytics).

Mr Bryn Balcombe (Chief Strategy Officer, Roborace) said that motorsport always played a role in displaying and advancing technology. He explained that Roborace is working on the development of autonomous cars and makes developer teams compete for the best software solutions, which are in turn implemented into the cars.

According to Balcombe, humans and artificial intelligence (AI) will increasingly be working together and complement each other. The importance therein is in determining the strengths of humans and the strengths of AI and how they can be linked.

Balcombe mentioned that USD$ 100 billion have been invested in the development of autonomous cars and noted that these investments should not only go towards the increase of companies’ revenues. According to Balcombe, some of that money must result in positive repercussions by reducing road deaths.

He recognised that economies will be increasingly data-driven and said that if networked and autonomous fleets – which will generate massive amounts of data – are deployed in cities, some of that information should be shared with the local authorities.

Mr Max Cavazzini (EMEA Automotive & Manufacturing Lead, Amazon) highlighted that Amazon has been using AI for years and is developing its AI-based portfolio through services such as Amazon Go, Alexa, and PrimeAir.

Amazon’s Web Services (AWS) are also very active in the automotive sector and provide Original Equipment Manufacturers (OEMs) with the possibility to host their data flows on the AWS platform given its advantage in cloud computing.

According to Cavazzini, legislation, and liability issues should be based on consumer perspectives and not only focus on OEMs standpoint.

Ms Anne Mellano (VP of Operations and Co-founder, Bestmile) pointed out that the automotive industry is looking to build perfect autonomous vehicles. The founder of the Lausanne-based company explained that the transition to autonomous vehicles will still be very long and that the ground for this transition must be prepared. Bestmile therefore offers business to business (B2B) solutions for mobility service operators and advises its customers on fleet optimisation and how new mobility systems might benefit them.

She noted that in the future, autonomous cars operating on shared driving platforms will be able to provide optimised services and avoid empty rides.

She said that intermediary steps in order to achieve fully autonomous vehicles causes confusion, given that it makes it unclear whether the car is controlled by a human or the operating system. This also causes challenges for regulators, as responsibility and liability are difficult to determine in these cases.

Mr Philippe Huysmans (VP of Growth, Ridecell) explained that three disruptions for mobility are currently happening: shared, electric, autonomous. Shared driving provides a lot of data about car fleets in terms of maintenance and business analysis. The electrification of car fleets might require a higher entry cost but will be profitable in the long run given the lower maintenance costs. According to Huysmans, the third disruption, the arrival of autonomous cars, will happen gradually. Therein, legal and insurance questions will become important issues. He explained that especially when it comes to insurance companies, early movers have a lot to gain. Autonomous cars will involve a lower risk of accidents which might lead to less revenues for the insurers. Therefore, according to Huysmans, many insurance companies are realising that it is more beneficial for them to collaborate with autonomous car developers to get access to data in order to adapt their portfolios, rather than trying to slow down an inevitable evolution.

Huysmans also noted that some cities are even ahead of the start-up wave by banning fuel-engines and encouraging smart city developments. However, these trends will drive more development in the industry.

Mr Alain Kornhauser (Professor, Princeton University, USA) explained that according to him there are three categories of automated cars: safe driving cars, self-driving cars, and driverless cars. Safe driving cars use technologies such as speedometers and lane centring. However, these technologies have increased the prices of cars and are often sold as extra features. Self-driving cars allow the driver to take the hands off the wheel and provide the driver with comfort and convenience. Automated parking assistants fall within that category. The category of driverless cars is what Kornhauser referred to as ‘mobility machines’ which are privately owned (e.g. by mobility providers) and could greatly benefit people with reduced mobility. This technology would enable mobility in remote, low-density areas where it is not profitable to run ride-sharing or taxi services as well as in cities.

Kornhauser also cautioned against too much reliance on deep learning and AI given that some conclusions of the software cannot be understood nor explained by humans.

Mr Julien Masson (VP Sales, CloudCar) presented the cloud-based infotainment platform, CloudCar. The principle of the software is to provide customers with an intuitive interface in the car which will allow OEMs to obtain some information about their customers’ data. Previous attempts by OEMs to provide their brand-specific car interfaces have not proved successful and interoperability with widely-used applications such as iTunes and Spotify have become the norm. However, in plugging in your smartphone ‘ecosystem’ into a compatible car system, the OEMs do not obtain information which could enhance the user experience. CloudCar therefore provides a cloud-based and personalised service which adapts to the user, all the while letting users enjoy their preferred medium.

Masson explained that the challenge is to make predictions of user preferences in order to make the experience enjoyable and relevant.

Mr Holger Weiss (CEO and Founder, German Autolabs) explained that his work focuses on what happens inside the car. He noted that conversational interfaces are on the rise, especially in the automotive industries, and that there is a general trend for ‘de-appification’. Since AI and cloud computing have increased precisiness of voice recognition up to 97%, there is an increased use of digital, voice-based assistants. German Autolabs is thus developing a digital assistant to be used in cars, which is proactive and provides a holistic platform for cognitive assistance.

Weiss also mentioned that so far, new types of mobility such as the shared driving economy is not traffic congestion nor pollution.

Mr Tomaso Grossi (Business Development Manager​, Autonomous Driving, Tomtom) said that automated driving relies on maps, sensors, driving policies and actuators. He further explained that OEMs are increasingly incorporating maps into their systems given that they can help anticipating the road ahead, reduce computation power and that they work in all weather conditions. Additionally, incorporating the maps and mirroring them with the data obtained from the cars improves their accuracy and allows OEMs to further take that information into account for the development of autonomous cars.