Autonomous cars were seen as a moonshot only 2 years ago, something that people couldn’t quite believe would get off the ground. An article in the Guardian from February 2015 titled ‘Self-driving cars and elevators to space: what will come of Google's bold 'moonshots’?’ (https://www.theguardian.com/sustainable-business/2015/feb/19/google-x-glass-nest-makani-driverless-cars-bold-bets-tech) articulated this well, essentially saying that the chances of a self-driving car was about as likely as creating an elevator to space.
Today autonomous cars are not something seen as an impossibility, they are not some kind of 80’s dystopian future idea. Instead, it is a case of when, not if. In October 2016, Google’s self-driving cars had covered over 2 million miles. At the same time, Jacquelyn Miller, a Google Spokesperson said:
‘We just got rear-ended again yesterday while stopped at a stoplight in Mountain View. That's two incidents just in the last week where a driver rear-ended us while we were completely stopped at a light! So that brings the tally to 13 minor fender-benders in more than 1.8 million miles of autonomous and manual driving — and still, not once was the self-driving car the cause of the accident’
On average somebody driving in the US has some kind of crash every 165,000 miles, so Google’s 1.8 million miles with 13 minor crashes puts it slightly behind, as an average human driver would have only 11 in that time. However, there is yet to be any collision between a self-driving car and traditional car that has found the self-driving car at fault, which is impressive, and given the ‘rubbernecking’ that is likely to take place when drivers see a self-driving car for the first time, it is not surprising that more people are crashing into them.
These initial forays into self-driving testing is creating a huge amount of data too, with the amount of data being communicated from cars to the cloud predicted to rise to 10 exabytes per month in 2025, which represents a 10,000 fold increase on the amount communicated today. This creates huge potential for companies, but also some potential headaches.
One of the key elements, with autonomous and smart cars, is going to be ownership of data, access to data, use of data, and protection of data. Ownership of the data being created is a key element within this challenge, given that it looks likely that several car companies will be using software and sensors created by other companies to create their autonomous fleets. A prime example of this is Uber’s experiments with self-driving cars, with their technology being retrofitted to a Volvo SUV for testing. Although companies like Ford, GM, Renault-Nissan, and Daimler are current industry leaders in self-driving, other companies who have been late with their development may use software from companies like Google for their offerings. This would work in much the same way that cellphone companies like Samsung, HG, and Motorola create the phones, but rely on Google’s Android software to make them run. In this case, does Google own the data? Or is it the company who built the cars? Or is the customer who purchased the car?
Each will want to have the data themselves, Google to improve their systems, the automobile maker to see any issues with their cars, and the customer because they don’t want their data being shared.
Data security is also going to be a key issue too, with a huge amount of very sensitive information available within the data, from the journeys that people are taking, through to how people act when by themselves in the car. If, for instance, somebody were able to see where a car was left every day for a long period of time, it would make stealing it or its contents easy.
Similarly, dealing with that amount of information will be a challenge in itself, with technologies, servers, analytical tools, and databases needing to be almost completely rethought to deal with it. In 2017 the world is going to create more data than the previous 5000 years put together, most of which will be left untouched, companies today are if anything over-collecting, creating a situation where they have entire brackets of data that sit in databases and are never used. When it comes to the amount of data companies could potentially get from cars, there needs to be some narrowing down of what’s needed, which can only come from experimentation after collecting as much data as possible, which puts pressure on data collection and analysis capabilities, which will therefore need to be reinforced before this process can begin.
However, data within autonomous and connected cars has the potential to have a significant positive impact on the world.
To some extent we are seeing this already with the use of telemetric boxes being used by many auto insurance providers, which sends them accurate data on usage, giving a better indication of driving proficiency. Historically the premiums of drivers were dictated by basic information like gender and age - for instance, the average 17 year old in the UK in 2015 paid £972 compared to the average 56 year old who paid £277. There will undoubtedly be some 17 year olds who are safer drivers than some 56-year-olds, but are penalized because of their age or gender. Through the use of telemetric boxes, insurance companies can study the data of drivers offering them premiums based on personal proficiency over broad base assumptions.
Most of the work currently being done in this area is being done by the world’s largest auto manufacturers and with an increasing number of connected cars on the road, if a company sells 500,000 cars per year, they have a data pool of 500,000 to test longevity, resilience, reliability, and performance for every part of the car with a sensor. They can therefore have a better idea of how an engine runs over time, how long a particular element of the engine deteriorates over time, and even which parts of the driver’s instruments are more difficult to use than others. This translates to a better designed future model or simple modifications or repairs to particular cars at particular times to keep them on the road for longer.
With autonomous cars in particular, the data they collect and analyze will be considerably wider ranging than connected cars because they will be looking at everything around them, rather than only the things inside or directly touching them. It will mean that road conditions, layouts, and infrastructures can be analyzed considerably quicker than they would in traditional settings. For instance, according to AAA, pothole damage causes an estimated $3 billion of damage to cars every year and in one case in San Diego, a cyclist successfully sued the city for $235,000 after she was injured after hitting one. At present the only way for local governments can fix potholes is if they are called in by members of the public, which is a risky and inefficient system. With autonomous cars constantly scanning the road for potential obstructions, they will have the data to
The opportunity that data from self driving and connected cars offers the world is huge, from increased safety, through to more efficient and better designed cars. However, there are challenges that will come from this use of data, we just need to make sure that we have the capabilities and wherewithal to deal with it.