Check out how you can help fight COVID-19! https://www.herox.com/COVID19

How Machine Learning will Improve Automated Driving

Hands-free driving is quickly emerging from the pages of science fiction novels and becoming a reality. Recently, Tesla released a software update, installed wirelessly, that enabled the automatic driving feature. Although plenty of car companies have released similar software - for instance, cars that aid with parallel parking, Tesla’s is the most advanced software yet. One big secret behind their success: machine learning.

Tesla’s machine learning algorithms use mapping and sensor data gathered from all other Teslas currently on the road. That information is then sent into a centralized database where it is analyzed and used to improve the autopilot software.

The new Tesla update is also, of course, plagued with all sorts of bugs and not exactly recommended for regular use yet. For example, whether the road being travelled is rough or smooth can seriously affect the performance of the autopilot system. However, it won’t take long before data from all the other cars being driven across the world makes automated driving a more pleasant experience. Autopilot systems also have benefits for everyone else on the road. Automated driving is potentially much safer than relying on human drivers; they have better reaction time and drive more predictably than humans. It could also seriously increase productivity: for instance, commuters could spend their long drives getting ahead on the day’s work or even catching up on some much-needed sleep (although it may be a long time before people feel comfortable taking it that far).

The tremendous benefit of cognitive computing can also be seen in Google’s self-driving cars. Another company synonymous with innovation, Google has also begun developing a self-driving car learns every time it drives. Google’s cars use a virtual map to navigate the road, more sophisticated than Google Maps, and described as incredibly detailed “digitizations” that track the height and position of every single curb. The car’s sensors then collect data while driving to integrate into the maps. All of that information is then aggregated into an updated map relayed to all the cars. Google’s car, so far, is not available to the public and the company is unsure whether it will ever release the model that they have been testing. The software they have been using, however, will soon be seen in other manufacturer’s vehicles.

For many people, the idea of self-driving cars is simply anxiety-producing. Tesla CEO Elon Musk has been very public about his skepticism of artificial intelligence, joining other ahead-of-their-time thinkers like Stephen Hawking and Bill Gates. However, in an interview with Tech Insider, Musk explains that he only worries about artificial intelligence being used with malicious intent. It’s hard not to worry about what the future will bring, since it’s unlikely that cognitive computing researchers and developers will slow down anytime soon. Like the internet, machine learning presents plenty of opportunities to those who would use it for the wrong purposes, but also presents an endless number of opportunities to use it for good around the world.

Contribute to research and development of machine learning programs right here at HeroX by getting involved with our Cognitive Computing Challenge.

more like this
GOVERNMENT TECHNOLOGY
comments
Non-Profit & Social Impact
lululemon announces the Healthier Community Challenge on HeroX
What if you could experience an environment that improved your well-being just by spending time in it? What would this look like to you? lululemon is…
1 min read
HeroX Tips: Formulating Your Prize
Determining the award for your crowdsourcing project can be a bit overwhelming! Thankfully, HeroX is here to help! First, start by estimating how muc…
1 min read
HeroX Tips: Your Crowdsourcing Timeline
Why is the challenge timeline important in crowdsourcing, when is it best to launch a challenge, what are the milestones that determine how long the project should run. All that and more tips and tricks that will help you launch a great crowdsourcing project or challenge, read inside.
1 min read