Exposé: Google Waymo Technology
With the current rise of autonomous technology, it’s common for consumers to start wondering about the future of the automotive industry. The supposed autonomous Chrysler Portal is going into production, Fiat Chrysler Automobiles (FCA) invested $30 million into the Chelsea Proving Grounds to turn it into a testing facility for autonomous cars available for consumer retail, and the Waymo self-driving vehicle project continues to expand through Arizona with the purchase of over 60,000 Chrysler Pacifica Hybrid units from Chrysler. While the journey of Waymo through the United States continues to expand, maybe it’s time to do a little exposé on Waymo tech to learn about how it works and how it can improve busy streets in the future.
First, let’s go over some statistics, about both the Waymo and human drivers. Just last year (2017), the self-driving Waymo had already logged over 5 million miles on public roads. In 2017, modified Chrysler Pacifica Hybrid units, the very same that make up the Waymo fleet, drove over 600,000 miles with only 124 human-driver interactions, or 124 times the on-board driver had to interact. Driver interactions for safety reasons dropped to 0.2 per thousand miles, and there have been zero reports of crashes or accidents.
In the world, there were over 1 million deaths due to vehicle crashes in 2014, with over 30,000 of those deaths happening in the US. The worst part? 94-percent of those deaths were due to crashes that involved human error. 9,000 were due to speeding, nearly 10,000 were due to intoxication, over 3000 were simply due to mobile phone use, and almost 1000 due to driving while drowsy.
Think of how much of a difference could be made if any one of these drivers had a fully autonomous, or even a semi-autonomous vehicle. With the latter being a vehicle that can drive the vehicle for the driver in various conditions, the last two scenarios could be completely avoided with advanced level 4 technology. And for those who may have had too many, sleep in the backseat of a fully autonomous vehicle on the way home – we’re sure most of the traffic the Waymo ride hailing app in Phoenix, Arizona will experience will be a fair share of late nighters who drank just enough to trust a robot to drive. Speeding is really up to the driver, but one would hope if the majority of people on the road have a self-driving vehicle, no one will need to speed, because there will be no need for random clusters of traffic.
Check out these specs.
Pretty fancy stuff right? Not just the technology either, but the design. The original Google Waymo vehicle had a rounded front and rear fascia to maximize field of view for the sensors, and designed the interior cabin purely for comfort since the car is doing the driving. On top of the vehicle is a LiDAR system, radar, and cameras that constantly map out the scenario by detecting objects in all directions. A LiDar with radar and cameras are also installed into the exterior mirrors and front fascia for a more complete image. Then, all of this data is compiled by additional systems for computing, steering, and braking, and the vehicle is operated by an onboard computer designed and built specifically for self-driving.
How does all of this work? Well, “LiDAR” is short for “Light Detection and Ranging,” or a laser range-finding system that maps out an object’s surroundings with multiple lasers. Common LiDAR systems fire 64 lasers simultaneously, and are connected to sensors that relay information from every surface a laser lands on. Google has tweaked their systems and efforts by going beyond just a bucket on the roof of a car so that the onboard computer can process and build a more complete wireframe map of the world.
Not just designed to make a map of the world, the Waymo technology has been configured to recognize and react to real-life situations. The vehicle sensors and software have been programmed to detect pedestrians, cyclists, other vehicles, and road work as far as three football fields (480 feet) in 360-degrees of direction. Now with nine million miles driven, the Waymo software and computers have learned to predict the many scenarios all of those variables may create during everyday traffic. The sensors can even so much as recognize a motor or bicyclist extending their arm to illustrate a hand signal, that will then cause the vehicle to slow down and make space for the cyclist.
The Road So Far
Google started the Waymo self-driving car project back in 2009 when they first started developing self-driving technology. Some of us may remember hearing those humorous stories about early Waymo vehicles being pulled over for driving slowly during testing phases of the early fleet. Those were the first concepts, modified Toyota Prius vehicles set for a goal of driving 100-miles ten times uninterrupted.
Once Google succeeded, they continued to push and improve the fleet with testing on highways, letting Google employees use them on the weekends, and basically collecting more data. The next model was based on the Lexus RX450h and made up the new testing fleet. In 2012, more than 300,000 miles had been driven before taking the vehicle to more complex city streets. Does anyone remember the video of Steve Mahan from the Santa Clara Valley Blind Center taking a test drive and going to Taco Bell in a Waymo? It was unbelievable.
In 2015, Google designed their own vehicle to continue to test the self-driving car, just in case they entered the automotive industry on their own wheels. Nicknamed “Firefly”, the new Google Waymo vehicle had custom sensors, computers, steering and braking, but no steering wheel or pedals. It was a prototype vehicle, and Google tested it with a unique collection of participants when showing it to the public.
Come 2016, the official Waymo company was established as the self-driving car company and the name for the vehicle so far, although owned by Alphabet Inc., a Google subsidiary. In the same year, Google and Fiat Chrysler started to work together with the late Sergio Marchionne adamant about getting in on the ground floor of autonomous vehicle technology.
It just took a year before fully self-driving Chrysler Pacifica Hybrid minivans started to appear, modified by Waymo tech. These units made up the first fleet of the mass-production vehicle, complete with a fully-integrated hardware suite for full autonomy. Having tested the tech in Austin, Texas, the Waymo-modified Chrysler was used in Phoenix, Arizona as part of an Early Rider program for consumers interested in testing the self-driving minivan.
The Early Rider program was so successful, Waymo actually became the first self-driving ride-hailing service in Arizona. The service should get a respective app like Uber or Lyft by late 2018, early 2019, and has already expanded its partnership to work with other businesses. Attention shoppers – you can now order groceries online at Walmart.com and then get picked up by a self-driving Waymo Chrysler Pacifica Hybrid minivan when your order is ready. This is just the beginning.
Now, where are my autonomous take-out vehicles? If we can’t have flying cars, at least bring me that sweet Chinese food boat from The 5th Element or the Pizza truck from Black Mirror. What do you think about all of this? Ready for the future of driving?
Let us know your thoughts and join the discussion on Aventura CJDR social media.