Self-driving Vehicles Are Closer Than They Appear
Waymo is a fully autonomous car created by Google. Waymo One is a ride-hailing service that is currently running in the East Valley of Phoenix, Arizona. Waymo Via offers autonomous trucking and local delivery. Waymo is also the subject of the latest Revisionist History podcast episode called, I Love You Waymo. In the episode, host, Malcolm Gladwell, test rides a Waymo in Phoenix where “unseen algorithms and artificial intelligence are guiding them unerringly toward their destination”. That mention is about all he goes into about the technology driving them – I’ll get into how autonomous vehicles work, and the technology, in a bit. First, we’ll go into the main premise of the episode: to prove that people should not be scared of self-driving vehicles.
Autonomous vehicles are not scary
Gladwell says that many people are frightened at the idea of self-driving cars and very clearly states that they should not be. As he travels around Phoenix in the back of a Waymo, he describes the “zen like” experience and how respectable the car was. He even tries to trick Waymo by placing a beachball in its way and even acts like a “crazy” runner running alongside it. Funny enough, he manages to confuse Waymo with his unpredictable running pattern and instead of continuing to drive, Waymo plays it safe and just comes to a stop.
Gladwell points out that pedestrians and cyclists are terrified of drivers. The reason we do not see cyclists and pedestrians competing with traffic is that they are very afraid to – “drivers are nuts!”. Drivers are not rational and we often lose sight of this. To help illustrate this thought, he interviews two former professional cyclists, Lance Armstrong and Jonathon Vaughters. Both top cyclists are in agreement that they fear drivers. Vaughters describes what he feels are two types of drivers. The first is the distracted driver. This is the driver who is texting, reaching for something, arguing with passengers or anything but paying attention to their surroundings. The second is the angry driver. The angry driver does not think the cyclist should be on the road at all. The road is for cars and anything else is impeding their progress to get to their destination. Vaughters says that this type of driver views the cyclist as “subhuman”.
Gladwell points out that: “Waymo doesn’t have emotions, text, eat a burger and drive with its knees, doesn’t act like an idiot, doesn’t give you the finger. Waymo doesn’t do road rage. Waymo wants to live and let live.” I think this is a very good point. What I personally fear about autonomous vehicles is that they cannot think – for some reason I believe that someone who has free thought, and is driving, will be safer. I assume they are always paying attention and will not hit me with their car. In practice, there’s no way I actually feel this way – I cycle and run all the time and I agree with Gladwell, I am terrified of traffic and will go out of my way to avoid congested roads. It is true that a machine does not have emotions and will not get distracted. So how does this self-driving machine work and keep us safer?
How Waymo works
Google’s goal was to build a safe, cautious and defensive driver – the technology would be a fully autonomous driving system and allow people to be only passengers. The technology used to allow for this is LiDAR, radar and vision systems and supplemental sensors. All this tech works together to help the car understand what is going on around it. The vehicle has a 360 degree view and can respond to objects up to 500m away.
The LiDAR system beams out millions of laser pulses per second, 360 degrees around the car. This measures how long it takes for the pulses to reflect off of a surface and return to the vehicle. The car then uses this information with other real-time information and updated maps to make smart driving decisions.
The cameras and radar system also work with the LiDAR system to help Waymo understand its surroundings. Cameras provide information about objects, like colours, and help spot traffic lights, buses, construction zones, etc. Radar is useful for various weather conditions such as rain, fog and snow. Other sensors, like audio detection and a positioning system, are used to detect emergency vehicle sirens and the location of other self-driving cars or regular vehicles.
How does Waymo “think”?
To work successfully, Waymo needs to be able to answer 4 basic questions: where am I?; what is around me?; what will happen next?; and, what should I do?
The “where am I?” question is answered by its mapping system. Waymo has pre-built 3D maps that highlight curbs, sidewalks, crosswalks, stop sights, etc. It also has the local traffic laws and regulations programmed. This pre-built information is used with the information Waymo gathers as it drives around in real-time.
The system is designed to operate in Waymo’s approved geography or a “geo-fenced” area. The maps of one Waymo are cross-referencing its real-time sensor data with its 3D map. If a change is detected, such as a collision up ahead, Waymo can reroute itself. The system will alert the operations centre and other Waymo vehicles, so they can avoid the area too. This ensures maximum safety for passengers and anyone else on the road. Eventually the idea is to build a fully autonomous driving technology that can take people anywhere – not just with the pre-programmed geo-fenced area.
The other 3 questions are answered by continuous monitoring done by Waymo’s sensors – Waymo can “see” up to 300m away in every direction. The vehicle scans for objects around, like pedestrians, motorcycles or road work, and reads traffic lights or temporary stop signs. Waymo anticipates the behaviour of similar looking cyclists, pedestrians and motorcycles. It knows that they move at different speeds and may change direction suddenly. Through real-world experience, Waymo constantly learns and improves its understanding of how different road users behave and anticipates what they will do next.
Waymo takes all the information it has programmed, with the information gathered in real-time, to determine the direction, sped and steering maneuvers needed to drive along its route safely. Waymo has defensive driving behaviours built in, such as staying out of blind spots and leaving extra space for cyclists. Through constant monitoring, Waymo can respond quickly to changes on the road.
Self-driving cars are in our future
It is crazy to think that autonomous vehicles are already in use in some American cities. At most, they are currently being tested in Canada, and commercial and personal use seems far off. Many people are afraid of the concept of self-driving cars, but in reality, human drivers are much scarier. The more autonomous vehicles are in use, like Google’s Waymo, the more we will come to realize how it benefits our lives. Waymo has no emotions, cannot be distracted and “knows” much more about its surroundings then a human driver ever could. Self-driving cars have LiDAR, radar and camera sensors that constantly feedback information that aid in anticipating the movement around and the decisions to be made.
The “I Love You Waymo” episode of Revisionist History plays an excerpt from a Ted conference where a presenter talks about a future with autonomous vehicles:
We may be the last generation to own cars. Our kids and grandkids, they may never have to learn to drive. They may never have to worry about driving around looking for parking or speeding tickets or drunk driving. And the best part is, they could go their entire life without ever seeing a car accident.A quote from a Ted conference. Source: Pushkin.fm
It’s strange to hear this and also exciting. I think we should take back the road and enjoy walking, running and cycling without the fear of getting hurt. I like a future where we put an end to road rage and relax and enjoy the ride.