Self-driving Cars Will Fail Unless We Redesign Our Cities
Urban planning can overcome the shortcomings of self-driving tech
It’s been nearly a decade since the launch of Google’s Waymo and Tesla’s autopilot feature, and yet, self-driving cars are nowhere to be seen. The excitement about the potential for fully autonomous vehicles that existed several years ago has since waned.
Of course, Elon Musk continues to promise the launch of a fully autonomous Tesla every year.
Driving is a complex activity. It requires the ability to process multiple information, analyze distance and speed, make judgment calls, adapt to different environmental conditions, and sometimes, the ability to hurl expletives.
Even as autopilot technologies have seen great strides in recent years, developers have recognized the limits of artificial intelligence, especially when it comes to interacting with humans.
Humans are far too unpredictable for any algorithm to predict behavior and react in a quick and meaningful way.
If we want to fulfill our dream of taking a nap while being driven to work by a virtual chauffeur, we can’t rely on technology alone.
Redesigning Cities
We can build roads that work for human drivers, or roads that work for autonomous cars, not both.
Throughout the last century, we have designed our cities around cars. Wider roads, parking spots, traffic lights, signboards, gas stations — our human-driven wheeled boxes need a complete eco-system to be fully functional. So why do we assume our smart cars will be able to thrive in this chaotic mess with just a camera and lines of code?
It’s time to redesign our cities around autonomous vehicles.
Provide Dedicated Lanes
The primary goal should be to cut down interaction between humans and self driven cars.
The first step is to create dedicated lanes for self-driven vehicles just like we do for bicycles and public transport buses. Having dedicated lines will drastically reduce the number of different scenarios that self-driving algorithms need to deal with. This will ensure consistent behavior of the autonomous vehicles, at least on these defined routes.
Dedicated lanes can also save self-driven cars from the wrath of frustrated humans.
Realtime Data Sharing
Adaptability is a weak point for computers. A human driver can understand when there’s been an accident on the road and adapt immediately by switching lanes or driving around.
While GPS is great for navigation, it doesn’t always reflect changes in road conditions, like when a new divider or construction work has been set up. If an autonomous vehicle doesn’t have prior information, it could drive right into a barrier.
Government bodies need to design an information system in conjunction with private companies, where such data is fed into these vehicles in realtime.
Yay! More data for Google.
Car Insurance
Self-driving cars are most valuable when all the cars on the road are self-driving, allowing higher speeds and reducing distancing in between cars.
Assuming the safety of autonomous vehicles over human drivers is clearly established (which is going to hurt a lot of egos), governments could push people to adopt autonomous cars by charging higher insurance rates for manually driven ones.
Cities will need to force adoption of automous cars through policy.
Sensor Based Navigation
Signboards and traffic lights are great for humans who are amazing at identifying patterns.
Humans can look at a colon, a hyphen and a closed bracket and immediately recognise it as a similing face :-)
Computers would struggle to differentiate between a traffic signal from a signboard in Times Square’s neon filled visual maze.
Cities need to install Smart Sensors which can communicate with cars and translate traffic and navigation data without leaving any scope for misinterpretation.
Promote Smaller cars
One of the major benefits of autonomous cars is the reduction in the need to own a car. But no one wants to get into a cab with three other passengers and wait to be dropped off before reaching the destination. Although I’d thoroughly miss the opportunity to eavesdrop on phone calls of fellow UberPool-ers and gain meaningful insights into their dating life.
In order to promote cab sharing, cities need to promote smaller two-seater cars similar to the microcars of the Netherlands. The limited speed and lightweight design of microcars ensure the safety of both the driver and the pedestrian. This would greatly increase confidence in autonomous vehicles and improve adoption rates.
While scientists and programmers are hard at work, trying to make a driver-free world a reality, citizens, and governments too need to lend a helping hand to usher in the future.
