Driverless Vehicles

4 min readApr 19, 2018

Driverless vehicles are one of those classic “fields of the future” that we all dream about: being chauffeured around by a robot so that we reclaim many of our hours each week affixed to steering wheel so that instead we can work, relax, or otherwise be more productive.

Automotive autonomy is classified into levels:

Level 0: No Automation — e.g. typical car
Level 1: Driver Assistance — certain limited control of steering or vehicle speed e.g. adaptive cruise control
Level 2: Partial Automation — steer, accelerate, brake in certain circumstances e.g. tesla auto pilot
Level 3: Conditional Automation — most aspects of driving can be controled — but driver must be able to take over at any time
Level 4: High Automation — can operate without any human input but in select locations and conditions e.g. Google prototype car
Level 5: Full Automation — automatically drive as well (or better) than a human under any conditions. e.g. None Yet — but many are working on it!

Background

Many people thing of the DARPA grand challenges that took place between 2004–2007 as the launching point for driverless vehicles. However, check out this interesting timeline of intelligent transportation systems at the DOT website. You’ll notice is that innovations and advances in driverless vehicles, sensing, mapping have been taking place since the mid 1980’s — 30 years ago!

The early limitations were in sensors, mapping, computation power, and the corresponding algorithms and computer vision and system level algorithms.

The DARPA grand challenges did bring together many interesting newer technologies and capabilities including Lidar (the ubiquitous spinning laser device seen on top of driverless card that helps sense obstacles) as well as computer vision and machine learning.

In recent years the acceleration of artificial intelligence — and particularly deep learning and computer vision have led an extraordinarily level of excitement about the inevitability of fully autonomous vehicles.

Why it it exciting?

This doesn’t need much elaboration but the promise of driverless vehicles is vast and wide — ranging from:

safety: in the US alone there are 30k deaths / year due to vehicles
saving time:
reducing carbon footprint
reclaiming our roads and greenspaces

and much more!

Market Opportunity and Implications

While the opportunities are huge — the broader societal implications might be even bigger.

There are currently 2M people in the US that have jobs as trucking. Another several million as taxi and other drivers.

Auto related industries such as car repair, car service, car wash, parking lots, gas stations currently employ hundreds of thousands of other employees . All of these won’t go away — but you can imagine many such operations could be centralized and streamlined. If cars can fill up gas on their own (or simply plug-in somewhere) — they will plan and go to a central location and the concept of a “convenience store” will change by its very nature; especially if you can get most things you want delivered right to your door. We likely wouldn’t need parking lots, valets, parking attendants, etc.

Market Landscape

Pretty much all auto manufacturers — including leaders such as GM, Ford, Tesla etc. GM for example bought cruise for $1B and Ford followed by buying Argo AI for $1B and several other acquisitions.

Intel bought computer vision company MobileEye for $15B. Uber pays nearly $1B for Otto — the driverless trucking company.

Google’s self driving car company — WayMo — has been buying thousands of vehicles to soon launch a driverless vehicle service.

Many new entrants — like drive.ai, cruise.ai — according to crunchbase — there are 370+ companies with hundreds of millions, if not billions of dollars of funding going into the space.

Here is a landscape map of various players in the Autonomy EcoSystem.