Location data and visualization studio for mapping, analytics, and more.
A brief history of autonomous vehicles, which dates back nearly 100 years to the World's Fair.
Autonomous vehicles (AVs) are attracting a significant portion of the automotive industry’s talent and capital — those attracted to cars that could be safer, faster, and never stuck in traffic.
Improvements in technology over the past several decades, such as LIDAR, GPS, and the internet seem to be converging that way, while also increasing the number of distractions for a driver — podcasts, audiobooks, movies and phone calls have joined radios to compete for a driver’s attention.
But the idea of a car that drives itself has been floating around for a surprisingly long time. General Motors’ Futurama exhibit at World’s Fairs before and after World War II featured radio control towers along highways like the ones used for air traffic control.
The current interest in them, however, dates back to the 1980s when companies like Mercedes-Benz and various universities, sometimes funded by the Defense Advance Research Projects Agency (DARPA), began experimenting with using computer technology to automate driving. These experiments worked on eliminating the need for external guidance like radio control or electromagnetic guides buried in the street.
In the 90s, the United States Congress even directed the Department of Transportation to come up with an autonomous vehicle system, but it was canceled.
Then in the 2000s, DARPA funded more projects. Larger companies, like Google, got involved when they announced their "autonomobile" concept, based on studies demonstrating that driver error was a leading cause of crashes.
Ever since, a number of companies have announced interest in the field and have inched closer to reality.
BMW developed a car that could learn a course when driven and then recreate it, down to the precise pedal movements the driver used.
Tesla is currently a leader in autonomous vehicle technology, although it has been forced by safety events to stop advertising its autopilot feature as self-driving.
Uber and Lyft have shown interest in self-driving cars, although both have sold their divisions developing the technology.
All the major car manufacturers are now working on self-driving technology to a greater or lesser extent, along with Google subsidiary Waymo. Several technology companies are also developing it, with the aim of selling it to manufacturers.
One of the more notable ones is Aurora Innovation, which was founded by former employees of Google, Uber and Tesla and which bought Uber’s self-driving research.
But while boosters and CEOs continue to make pronouncements about the imminence of self-driving cars, many technical and legal challenges remain.
All self-driving technologies on the market still require a human operator to take control in some situations, such as construction zones, when there's incomplete location data, and sometimes in the presence of cyclists, pedestrians and other cars.
Morning Brew’s Ryan Duffy’s encounter with Waymo One, a driverless taxi service, last July exemplifies the current state of the technology.
Vehicles were confined to a service area in the suburbs of Phoenix, AZ and had very strict parameters about where riders could be picked up and dropped off. The cars didn’t go on the highways and tried to avoid traffic; they also exhibited occasionally erratic braking behavior, which area police blamed for a few minor collisions (although Waymo blamed them on human drivers).
But the vehicles still have problems navigating through anything that hasn’t been mapped thoroughly or when dealing with unexpected situations, such as cars parked illegally, construction zones, and even fireworks carts set up in parking lots for the Fourth of July.
Development will continue and no one is sure when, or if, fully self-driving cars will be available.
With current iterations, one can imagine large scale deployments serving an area between an airport and a convention center, or on a corporate or university campus.
One company, Motional, intends to launch services in multiple cities in 2023 in a partnership with Lyft, according to TechCrunch. However, all of its vehicles will have a human present.
Most of the work on driverless cars performed in real world conditions have been in pedestrian-unfriendly SunBelt cities, like Phoenix and Las Vegas, or San Francisco — places with very consistent weather. Successfully navigating Boston, with its changeable weather, unpredictable drivers, multitudes of walkers and cyclists, is likely to be a great challenge.
The ongoing semiconductor shortage may also increase costs and cause delays in any large scale deployment for the next few years.
More existentially, autonomous vehicle developers have to wrestle with the fact that AVs have limits. They do not reduce congestion, for example, they merely replace it. Similarly, wear of tires can result in even fully electric autonomous vehicles being a source of air pollution.
But cars are not the only autonomous vehicles.
Related to the efforts to develop self-driving cars are efforts to make self-driving buses.
These are more advanced than cars because they can follow a fixed route and separated guide way, greatly simplifying and reducing the complexity of self-driving.
Similarly, autonomous trains have regularly been used in metro systems around the world for many decades.
Labor costs are among the highest costs for many transit agencies, so automation would help them save money and provide a more accurate service.
This is to say that autonomous transportation does exist, and has found success in certain sectors. It is much more likely that we find ubiquity in autonomous vehicles in the transit space before we do in automobiles.
And while driverless cars may one take you from location to location, there are still many gains to be made before driverless cars flood your streets and highways.