Building the Operating System for the future of transportation

In 1900, Americans owned just 8 thousand cars and the horse-drawn carriage was the country’s primary mode of local transit. By 1920, there were 8 million cars across the U.S. and, thanks to Henry Ford’s mass-produced Model T, the carriage industry was well on its way to irrelevance.

Today, we are once again in the midst of a once-in-a-century transformation of our primary mode of transit. Whereas in the early 1900s we reinvented how vehicles were powered, replacing the horse with the internal combustion engine, we’re now rapidly reimagining the way our cars are controlled, with the human driver on the verge of being displaced by the robot.

The outcome of what is shaping up to be a complete disruption of the $5 trillion global transportation industry seems inevitable: the replacement of more than 1.2 billion private cars and buses with electric, autonomous vehicles that are summoned on-demand, and -- if these vehicles are to reach their full potential for providing low-cost, efficient rides -- are shared by multiple passengers.

To realize this future of autonomous, on-demand “shuttles”, we must first solve a number of key technological challenges, not least of which is enabling cars to safely drive themselves. Solving the problem of autonomous driving has attracted some of the brightest minds in the world, but what is often forgotten in the process is a second, equally challenging problem: creating the software that coordinates, in real-time, the movement of millions of connected, human-driven or autonomous shuttles and their passengers.

At Via, we refer to this software layer, which dynamically routes vehicles and matches them with passengers requesting rides, as the “Operating System” for on-demand shuttles. We have been hard at work building this Shuttle Operating System for nearly five years.

To operate efficiently, the Shuttle OS needs to: (1) automatically identify riders whose routes overlap, so they can share a single shuttle without creating any detours; (2) match those riders with the most suitable nearby shuttle; and (3) direct that shuttle along the fastest route to pick up and drop off its passengers. Solving this problem across thousands of vehicles and tens of thousands of riders, while also coordinating the seamless pickup and dropoff of each passenger, is a major computational and operational challenge that requires sophisticated algorithms and tremendous amounts of data.

We first deployed our Shuttle OS in New York City in September 2013, and it has since powered more than 10 million efficient, convenient, and affordable shared rides across NYC, Chicago, and Washington DC. During this time, our algorithms have improved tremendously, continuously learning from the billions of data points we’ve collected. Our software has learned to intelligently select the fastest routes for each shuttle using real-time traffic data from multiple sources; direct passengers and shuttles to optimal pickup locations to avoid detours; and accurately predict how demand patterns will unfold throughout the day and smartly position vehicles to anticipate that demand.

While the ultimate application for our technology may be the on-demand, autonomous shuttle of the future, the system we’ve developed is also highly relevant to so many of today’s urban transportation modes. Our Shuttle OS can be used to operate a new form of public transit: a fully dynamic network of shared vehicles, free of the constraints of fixed routes and fixed schedules that limit the efficiency of traditional bus systems. It can be used to dramatically increase seat utilization in traditional taxi systems, enabling shared taxi rides. It can provide a shared mode to car and taxi e-hail apps, powering real-time urban carpooling. And it can be applied across college and corporate campuses to provide an efficient local shuttle service.

Our goal is simple: to make our OS available to every transportation operator, public or private, across the globe. We are actively partnering with public transit authorities and municipalities, with taxi companies, private transit contractors, and transportation network companies, and with auto manufacturers and developers of autonomous vehicles.

Today, we’re excited to announce the launch of our first commercial partnership: an on-demand shared ride service in Paris, the result of a close collaboration with Keolis, one of the world’s largest public transit operators, and LeCab, a fast-growing French on-demand mobility provider. The official press release is here.

Stay tuned: we’ve got so many more partnerships in the works that we can’t wait to share. Until then, if you happen to be in Paris, check out our new service!

-Oren Shoval and Daniel Ramot, Via Co-founders