Follow Datanami:
August 16, 2019

Learning from Data to Create a Safer and Smarter Self-Driving Experience

Pete Brey

(metamorworks/Shutterstock)

The automotive industry isn’t just being driven by people — it’s also driven by data, particularly as automobile manufacturers move toward autonomous, self-driving vehicles. Last year, Waymo cars drove 1.2 million miles in California. Meanwhile, Tesla, with its Autopilot program, is actively collecting data from hundreds of thousands of vehicles to predict how its cars might perform autonomously. So far the company has collected hundreds of millions of miles worth of data.

What are these autonomous vehicle manufacturers doing with all of that data? How is it being used to help them make intelligent decisions that can help make autonomous driving safer and more prevalent? And how can they make sense of it all to produce a better and more reliable driving experience?

Stuck in the Data Lake?

Like other industries, many automotive manufacturers have embraced the concept of data lakes as a means of housing massive amounts of information. Data lakes enable organizations to store content in various formats, including structured and unstructured data. That’s especially important for autonomous vehicles, which generate data streams in different formats depending on the information that’s being collected (i.e., the data from a camera will be unique, as will the information derived from the car’s performance in certain weather conditions). Data lakes provide the scalability and storage that enable the automotive industry to address this important market.

The challenge is that many of the data lakes that automobile manufacturers have deployed are siloed, while others run the risk of becoming data swamps. Manufacturers often collect many copies and tiers of data, all of which needs to be processed for actionable insights. They need to be able to take raw and “dark data” — data that is collected but not effectively analyzed or processed — and use artificial intelligence (AI), machine learning (ML), and deep learning (DL) to improve the performance of their autonomous vehicles over time.

(ZinetroN/Shutterstock)

Driving Intelligence Through the Lake

Deploying AI, ML and DL processes in conjunction with data lakes can help companies effectively cut through the data noise to determine which data points are important now and which can be saved for later. These technologies can harness incoming information and provide recommendations and intelligence in real-time. Manufacturers could better understand how a vehicle performs in heavy traffic or inclement weather, for example, allowing them to make changes to a car’s system if necessary.

This data can and should be stored for future analysis. AI and ML evolve based on historical data and can be used to provide better insights based on past precedents. As car manufacturers collect greater amounts of data, they’ll need an infrastructure that allows their analytics systems to upload and access that data as necessary, no matter where it is. Open software-defined storage, which can be massively scalable and used with industry-standard hardware infrastructure, can accommodate manufacturers’ ever-growing data management requirements.

Data and the Drive for Safety and Reliability

While many industries have turned to modern data storage and analytics solutions to address their data management needs, the automotive industry’s requirements are different than, for example, retail or telecommunications. Autonomous vehicle manufacturers want to deliver a great customer experience, but right now their primary focus is on safety and reliability to better equip cars.

Data can and is helping, but manufacturers need the correct infrastructure in place to manage the information that’s being collected. That infrastructure should be highly scalable and accompanied by advanced AI, ML, and DL capabilities. This winning combination can enable manufacturers to glean insights that can be used to put a better product on the road.

About the Author: Pete Brey is marketing manager of hybrid cloud object storage at Red Hat, including Red Hat Ceph Storage and Red Hat data analytics infrastructure solution.

Related Items:

AI Partners Shift Driverless Cars into High Gear

Bright Skies, Black Boxes, and AI

Ford Accelerates Driverless Car Effort With Machine Learning

Do NOT follow this link or you will be banned from the site!
Share This