Despite its early promise, autonomous driving has ended up on a bumpy road. Simply put, it’s hard to train software to handle the chaos and unpredictability of the modern road—especially if you’re trying to avoid killing anyone. Nonetheless, Uber is reportedly giving its self-driving efforts a reinvigorated try, this time in Washington, D.C.
“Starting on January 24, we will begin exploring a new mode of transportation in Washington, D.C. when we bring our self-driving cars to District roads for data collection,” Danielle Burr, Head of Uber Federal Affairs, stated in a Medium posting. “Importantly, these cars will be in manual driving mode一meaning a Mission Specialist (a specially trained vehicle operator) will maintain control of the vehicle at all times.”
In other words, Uber is a little more cautious this time around. That makes sense, since one of the company’s autonomous cars killed a pedestrian in Tempe, Arizona in 2018. In Washington, D.C., Uber’s phased approach will involve manual data collection, which in turn will allow the company to (in Burr’s words):
“Develop high-definition maps, the foundational information layer for our self-driving system.
“Capture scenarios that might be different from city to city to simulate and incorporate into on-track testing.
“Further refine our expansion methodology, which involves identifying key characteristics within an operational domain in a new city, and then running that data through our autonomy system in simulation and on our test track to verify that our self-driving system performs as intended before it drives on the road.”
Uber has spent enormous resources (and lots of years) on developing autonomous driving, which it sees as key to its eventual profitability. It’s also facing quite a bit of competition: Waymo, Google’s autonomous-driving spinoff, is testing self-driving taxis in Phoenix and Silicon Valley. Tesla is likewise dedicated to self-driving vehicles, with CEO Elon Musk recently predicting that the company’s automobiles will soon hit full “Level 5” autonomy. Then there’s Apple, which is reportedly doing… something… with automobile technology.
Autonomous Driving: Still a Career Option
So, despite the difficulties, companies seem determined to evolve the technology that underlies autonomous driving. And that means it’s still a viable career pathway for technologists who want to change how folks operate on the roads.
If you’re totally new to autonomous driving as a concept, Udacity offers a self-driving car engineer nanodegree, and Udemy and Coursera have similar coursework. If you’re interested in a more academic approach, MIT has videos and slides from its “Deep Learning for Self-Driving Cars” coursework available for your perusal.
Working with autonomous driving will also require knowledge of everything from computer vision to machine learning. These are advanced concepts that require quite a bit of specialized study; if you’re looking for an extensive, more generalized introduction to machine learning, Google has a (free) course with 25 lessons and more than 40 exercises, which you can finish in roughly 15 hours (and features lots of interesting video). There’s also Bloomberg’s Foundations of Machine Learning for technologists with strong experience in machine learning and mathematics. But that should be just the beginning of your learning.
Because of the money needed to develop a commercial product, and the amount of knowledge required to actually build the necessary software and hardware, autonomous driving is also likely to remain a relatively small endeavor for the next few years, in terms of employee count. But that doesn’t mean someone with the right skills—and sufficient motivation—can’t land a job in it. Despite the bumpy road, the industry is still committed to this route.