Will self-driving cars ever be able to safely drive themselves?

Auto / Tech

There have been numerous reported accidents involving self-driving cars, but aside from that, we have to ask ourselves whether or not this technology can fully integrate into our lifestyles.

Automated technology has grown leaps and bounds over the last ten years providing endless applications for individuals and companies around the world. From our personal lives with AI assistants and tailored media, to the business world with manufacturing and communication, automation is here to stay.

While a large group of automated applications have already begun to increase efficiency, there are still subsets that have not quite earned our trust. Even though we may be suspicious of our personal assistants always listening to us, one specific focus area that has a long way to go is the safety of automated vehicles.

Self-driving cars have been around since Google announced the Self-Driving Car project back in 2008, and while this realm of technology has improved immensely over the last decade, continual issues plague any notions we may have of widespread adoption. There have been numerous reported accidents involving self-driving cars, but aside from that, we have to ask ourselves whether or not this technology can fully integrate into our lifestyles. For example, how are self-driving cars to navigate automatic car washes?

Back to the safety aspect, companies like Tesla and Uber have been in the spotlight recently because of accidents involving their self-driving cars. However, for the sake of consumers who may eventually decide to purchase or go for a ride-along in one of these vehicles, there are companies who are largely invested in improving the technology.

One such company is an unassuming camera technology company called Light. Based in Palo Alto, California, the team at Light is familiar with the self-driving car market. Not only this, but their technology has positioned them to expand into additional B2B solutions, specifically advanced vision for automated vehicles. While their L16 computational imaging camera already brought 16 lenses to consumer photography, the company plans to use its lens array to improve some of the automated vehicle technology already working its way onto our roads and highways.

self-driving car
Prototype of the Google Self-Driving Car

Briefly, technology like this works similarly to the human brain. Multiple pictures are taken and then stitched together in order to create a cohesive and easily understood image. A lens array and software work in tandem to allow machines to navigate and understand three-dimensional spaces.

From a deeper technological perspective, appropriately calibrating an imaging system’s depth resolution is extremely important when it comes to getting automated vehicles safe enough for regular real-world use. Depth resolution will vary depending on the type of application the system is intended to be used for. Currently, advanced driver-assistance systems (ADAS) like forward collision warnings and adaptive cruise control utilize fine depth resolution, whereas advanced robotics require ultra-fine depth resolution to complete complex tasks. Considering that the ADAS market is expected to reach nearly $70 billion by 2024, it is safe to say that companies will be investing heavily in this arena.

Real-world applications require a heightened level of accuracy and efficiency, especially when it comes to getting automated vehicles safe enough for regular consumer use. With imaging systems improving, the software that works in tandem needs to improve as well. Luckily for consumers, research teams have been working diligently to improve the artificial intelligence capabilities when it comes to driving on our roads.

Now, some people may scoff at the idea of using video games to teach self-driving cars how to drive safely, but that is exactly what is happening. The fifth iteration of Rockstar’s franchise behemoth, Grand Theft Auto, is being incorporated into the research and development of advanced imaging systems for self-driving cars. Research teams have been able to develop software that is able to effectively recognize various objects within the game like pedestrians, other vehicles, and curbs in real-time. Because video games like Grand Theft Auto 5 push the limits of realism with their graphics, the environmental data scraped from the game is just as good as the data would be from real-world traffic scenarios.

Obviously, we are far from living in a world where self-driving cars outnumber the typical cars of today, but there are promising advancements in technology and upcoming trends for 2019 that show we are well on our way. There are also many unknown factors at play here, like whether or not consumers will trust the technology or if our infrastructure will need to be adjusted to accommodate. Despite all of the unknowns, for every safe road test that occurs with an automated vehicle, we are one tiny step closer to it being a reality.

Do you think we will ever get there? Would you ever take a ride from a self-driving car, or better yet, purchase one? Let us know in the comments below or on Google+, Twitter, or Facebook.

Comments
To Top