How do autonomous vehicles work?

What in the Bumblebee (yellow Chevrolet Camaro) story is going on with these vehicles? No hands, no feet, not a corporeal human presence needed. Let’s begin our exploration.

Humans and digital technology go way back. Humans and vehicles go even further back. Behold now, this relationship is so at peak that vehicles can handle digital technology smoothly than its master drivers.

Modern romantic

The invention of the autonomous vehicles (AVs) technology marks a pivotal point in the trajectory of the automobile and transport conglomerate. Every new invention in the technological field of vision, especially those that seem utopian into-the-future, are deemed a sophisticated fancy; modern and romantic in nature. Same goes for AVs.

An easier adjective for these vehicles is ‘self-driving’ which is quite apt for how it presents itself. An autonomous vehicle is one that can drive without assistance from the driver. Yet, how do these vehicles “self-drive”? They clearly aren’t advanced robots from Cybertron.

Role number one – Vision

Driverless vehicles work effectively due to the fact that they can see. Sounds like some fantasy come true, doesn’t it? What actually is meant here is that there are many radars and sensors installed within the vehicle that can detect and ‘see’ what’s going on around it. In a Waymo car (a self-driving car by Waymo LLC), there are LiDar sensors which the car heavily depends upon to see. LiDar is essentially light detection and ranging where data collected creates a 3D map of environments for the vehicle. The picture of the Waymo here shows you the roof-mounted LiDar sensors.

Role number two – Monitor

Since an active vehicle means it’s on the move, its environment will change every second. It’s crucial for AVs that they constantly and constantly monitor the data that is fed to them. The cameras, the LiDar sensors, radars – all provide data of what’s happening on the road. Whether someone is crossing the road, if there’s an upcoming speedbreaker, or a high kerb, if there’s a stop sign, and much more. The radars will calculate the distances to all these.

The spinning LiDar on the roof renders real-time 3D images, for example like this one here, of the people all around.
| Photo Credit:
Steve Jurvetson/Flickr

Role number three – Control

The data is processed by the software (artificial intelligence works its hand here) and processors which in turn signals other parts of the vehicle to do the required actions — such as braking, slowing, steering, stopping, etc. The vehicle gathers the most responsibility here — control and action. Artificial intelligence closely works with the sensors and radars in the system to make decisions after perception. Whether the car should brake now or speed up or even overtake; such decisions will be taken by the car.

Then again, it’s not all unicorns and rainbows. Our streets are filled with all sorts of things and oddly, a primary limitation is that these vehicles are not driven by humans and hence lack sentient awareness on the road. The sudden jolts and shocks while riding AVs reveal the many facets that need improvement.