How Data Managers are Steering Us Toward a Better and Safer Future on the Roads

Autonomous vehicles are on the rise to combat the country’s motor vehicle fatalities. This article by Red Hat’s Pete Brey takes a dive on how machine learning, artificial intelligence, and deep learning work together to achieve this goal.

Houston, we have a problem. So does Los Angeles, Atlanta, New York, D.C, Boston, and all cities, towns, and counties throughout the United States. That problem is motor vehicle fatalities. According to the National Safety CouncilOpens a new window , there have been three straight years of at least 40,000 roadway deaths in the U.S.

Accidents aren’t the only things plaguing U.S. drivers and passengers. Traffic is also getting out of hand. Traffic congestion cost the U.S. economy nearly $87 billionOpens a new window in 2018, thanks in large part due to lost productivity. That’s not even accounting for the psychological damage that can be caused by sitting in a traffic jam for hours.

The race to address the issues of road safety and traffic congestion has led to a fascinating marriage between the automotive and technology industries. Understanding that most accidents are caused by driver error, many auto manufacturers are diving headfirst into autonomous vehicles (AVs).

Google’s WaymoOpens a new window AVs drove 1.2 million miles in California in 2018, and Tesla is moving full speed ahead with its AV plans, with the goal of having “one million robotaxis”Opens a new window on the road within the next year.

Contrary to popular belief, these cars won’t be driving themselves. They’ll be driven by data, along with artificial intelligence (AI), machine learning (ML), and deep learning (DL): three technologies that are closely related, yet different.

AI essentially translates human intelligence to computers; ML empowers those computers to use that intelligence to process information and “learn;” and DL involves identifying patterns, classifying information, and comparing that information to previously known data. All three are essential to AVs.

Imagine a line of self-driving cars, guided by millions of data points, that can navigate roadways without gridlock or make split-second decisions to avoid collisions. This is what car manufacturers and technology companies are driving toward–and data managers are in the driver’s seat. Through use of the aforementioned technologies, managers have the power to make getting point A to point B safer and smoother.

Here’s how they’ll do it.

Getting a feel for the road through AI, ML, and DL

AVs are equipped with several sensors, including cameras, radar, and lidar. These tools can collect massive amounts of data that can be used to calculate everything from a car’s performance under certain weather conditions to its ability to avoid collisions.

Applying AI, ML, and DL to this raw data can help manufacturers infer actionable decisions they can use to improve their AVs. They can identify which data points are critical for current performance standards and what those points mean. From this information, they can derive intelligence regarding what adjustments need to be made to optimize the driving experience.

AI, ML, and DL can also identify data that may not be critical today, but may be important in the future. Historical data can be used to learn about how an AV reacted under certain situations (“this happened here, the car did this, it should have done that”).

ML and DL can analyze these patterns, effectively “learn” from the mistakes the car may have made, and provide intelligent recommendations on how to address the errors. The more data the system collects and stores, the more it learns, and the better the AV can become at being able to adapt and provide better, safer, and smoother transportation.

Changing the way data is managed for the benefit of society

While the information is being used to help the cars make real-time decisions, it’s still up to data managers to make sense of the information–essentially telling cars when and where to go, when to stop, and how to react in a given situation. However, that can be challenging given the sheer amount of data that these cars have the potential to collect.

This could require data managers to change the way they manage and store information. AV manufacturers use large-scale data lakes to siphon and collect information on their cars, and they need storage solutions that can scale well into the petabyte range and perhaps beyond. Open source software-defined storage can be an ideal option, thanks to its high degree of scalability and ability to work with any hardware infrastructure.

The people in charge of managing and interpreting that data are today’s Henry Fords. While Ford didn’t invent the automobile, he did revolutionize the business through the introduction of the Model T. Now, it’s data managers’ turn to take the wheel and steer us toward a better and safer future on the roads.