Garry Golden

[Update on 1/21/08 -  Since writing this post, General Motors declared  that commercial autonomous vehicles are ‘closer than most people assume’; India won cheers and jeers with its $2,500 Nano ‘people’s car‘, and the web has been active in software oriented transportation solutions that tap GPS devices jumping off the shelf at Best Buy;  Peer to peer networks, IBM’s vision of five innovations including vehicle to vehicle communication systems.]

Want a transportation problem that could eventually become a bigger headache than oil?

Try congestion - and the fundamental performance limits of human operators!

Productivity losses, the emotional baggage of commuting, missed time with family, losses related to freight, wasted energy, expenses of adding road capacity, etc. We could also expand congestion to one of its major sources – accidents caused by human error, and calculate the economic costs resulting from injuries and fatalities.

Want a date for the history books of transportation?
November 3, 2007 - when three fully autonomous ‘robotic’ cars crossed the finish line after driving along a 60-mile ‘urban’ course without human instruction or guidance. These vehicles responded to real world simulated conditions and made decisions to changing events on an abandoned military base. The winning vehicles were designed by college professors and engineering students competing in the latest variation of (US Military) DARPA’s Grand Challenge. (Video below)

The idea of a ‘driverless car’ is not new. But we are now ready to have a more enlightened conversation that knocks down the walls created by emotional responses to a big 21st century idea - that vehicle based computer systems might be a better platform for the transportation of goods/people than human operators.

We are certainly at the point where it makes business sense to invest in commercializing these systems. And if you framed the issues broadly enough, autonomous vehicles could easily become a state or national policy issue in the years ahead -

In 1990 most people had no use for a ‘home computer’ - no need for a network connection that went beyond voice data. By 2000, the digital divide/’broadband’ access became a national issue of economic competiveness.)

Can we imagine political support for autonomous freight vehicles to keep major trade routes competitive? Or a region’s high adoption rate of autonomous passenger vehicles changing where major employees locate their businesses? When might the President say the words ‘American development of autonomous vehicles’ in a State of the Union Address?

The evolution into autonomous systems is not the same type of ‘leap’ into flying cars. It is real and very attainable given existing infrastructure and trajectory paths of technologies. The biggest uncertainty is the human/societal response. (Hence, my call for a more enlightened conversation.)

Back to DARPA….
Techno-enthusiasts & engineering devotees have been overjoyed with the progress over the past five years of Grand Challenge events. Following the success of the Urban Challenge, most major media outlets treated the event with a snarky ‘futuristic’ tone… Isn’t it scary to think about robots driving on the road. I don’t like that idea…

But for anyone under 40, it might be time to have a more enlightened conversation about the future of mobility. And for entrepreneurs, it is time to consider the tremendous wealth creation of ‘robotic’ autonomous vehicles.

Yes, we are still years away from seeing fully autonomous vehicles on public highways, but we now have a clear picture of the path (and profits) ahead. And the obstacles look like speed bumps.

Why accelerate the development of autonomous vehicles?
To remain economically competitive by improving productivity and generating new wealth systems.
The number of vehicles on the road (globally) is set to explode over the next fifty years – and we might need to fundamentally change our approach to the movement of humans and freight. Increasing traffic flows through autonomous systems is a perfectly reasonable option to explore. I might even say, inevitable.

More importantly, we must also look beyond the impact of autonomous systems on personal travel to consider the market for autonomous vehicles for freight, agriculture, mining, energy harvesting, public/private mass transportation (e.g. rapid bus transit), and applications that we cannot currently imagine.

Autonomous systems are likely to go far beyond our ‘highways’. Consider the role of much smaller (’robot sized’) autonomous carts operating in warehouses, hospitals, factories and offices adding tremendous value to back-end operations.

How many human operated vehicles should we expect by 2050?
A recent NY Times article quoted Ford Chairman Clay Ford Jr. predicting‘…the world’s automotive fleet, now about 880 million vehicles, would grow to 2 billion by mid century.’ [and that] ‘congestion is going to be a big a problem as pollution.’ [And, if I am correct, that does not include two wheeled vehicles.]

There is a cost to supporting these vehicles under the assumption that human drivers will determine the capacity limits of roadways. Balancing strategic investments in autonomous systems might be a more cost-effective way of absorbing more vehicles onto the world’s road networks. It would also require starting enlightened conversations about what is possible within the transportation sector over the next thirty years. (And which directions we’d like to support…)

Autonomous systems, in theory, can radically expand the number of vehicles ‘front to back’ along a road. Imagine a virtual ‘train’ of cars only a few feet apart, working in unison, in constant communication and collectively monitoring all external road conditions. One disruption along the chain of vehicles and the entire system reacts instantly. These systems could also increase the number of vehicles ‘left-to-right’ – now imagine smaller, slimmer, smarter vehicles packed side-by-side in a lane that (today) only holds one vehicle. [And no, I don’t believe in a future for flying cars! So we’ll save the ‘third dimension space’ for another post on regional air travel!]

Why do I believe this problem will appear sooner rather than later?
1) Profit potential / Market pressures - The automobile industry stands to make much more money from developing ‘smart’ vehicles then it might with ‘green’ technologies. I doubt that the ‘post 1950s era’ business model of simply building and selling cars is the platform for growth in the next century. Real money might be made around after market ‘add ons’ and in-vehicle digital services.

2) Making customers happy - Looking at the global market for vehicles it is obvious - ‘small’ is the new ‘big’! Market growth is happening in China, India and elsewhere. The first step into the marketplace? Small, cheap vehicles.

In recent months, there have been widely published media reports about India’s plans for a $2,500 car, China’s plans to sell $7,000 cars in Mexico, Nissan’s plans for a $3,000, et al. The list of brand names goes on and on. But all share the same plan. Build lots of inexpensive cars and jump start global auto sales. (Hence, the ‘oil’ supply problem!!)

It is not difficult to extrapolate this strategy of flooding ‘cheap cars into India/China’s middle class’ out into a nightmare scenario of mega-city congestion.

Looking beyond that first wave of buyers (2010-2025), the auto industry might need a new type of vehicle to keep growth rates going in Asia. ‘Small, cheap and smart’ is a better combination.

A very relevant question is - how might Asian/South Asian cultures respond to markets of autonomous robotic vehicles? Is anyone asking…?

So what about the ‘Robotic’ vehicle race?
DARPA’s Urban Grand Challenge is designed to meet the U.S. military’s goal to replace 1/3^rd of its ‘operational ground vehicles’ be unmanned by 2015.’ (Eight years from now..!!) It has run variations of the Grand Challenge since 2003 and progress has been quite impressive. In less than a decade from now, billions of dollars will be spent on advancing autonomous systems currently being constructed by Professors and Graduate students around the nation.

Putting all the pieces together
The software/hardware components of autonomous systems are coming together quickly. And I cannot see any major barriers. Will it be challenging? Of course - especially around system integration issues! But there is no major component missing from the list that is not currently in developement.

The roadmaps for sensors seem achievable and we can expect breakthroughs in algorithms and integration systems. I suspect we’ll see significant progress over the next 10-20 years. The major barriers will naturally come from the human sphere - insurers, policy makers, transportation agencies, and consumers.

Confronting Congestion with Lightning McQueen™
Yes, there are hurdles ahead with adoption. Yes, I expect humans will always want to have the ‘drive option’ - when roads aren’t congested or it’s just too fun to drive on country roads.

And yes, today the idea of ‘car driving driver’ is a painful, scary concept to accept. But the conditions which shape our attitudes towards autonomous vehicle operations will evolve. And it would be foolish to shape a forecast based on current mainstream opinions and worldviews. This is a future for those born after 1980.

 

 

Things will change slowly. It will take time and evolve over stages. Most cars already have alert systems for avoiding backing up into small children/fixed obstacles. By 2015, I suspect semi-autonomous ‘adaptive cruise control’ will be standard in most vehicles. And then we’ll be able to have a more fruitful conversation about the benefits of fully autonomous vehicles. I would expect that the first autonomous systems will NOT be on highways. So the first wave of commercial products should occur in more industrial environments. (warehouses, construction sites, mines, airports)

Who are the players to watch?
Following the event, the NY Times reported “…progress has been so dramatic that the impact is likely to be felt soon and far more broadly, in the commercial automotive world and elsewhere.” Remember that the DARPA contest has been driven by University teams, not major industrial giants.

Autonomous vehicles are really systems of systems. Standards are being established (e.g. (OSEK/VDX), and Japanese companies are calling for standards to push the industry forward. Otherwise the playing field is wide open and revolutionary products are being developed by companies that do not have household names.

A recent Forbes article, takes a more conservative approach to the pace of development, but highlights several companies in position to profit: Cognex (Natick, MA), IBEO/SICK (Germany). I am looking more closely at component builders - Trimble, Evolution Robotics, Sensata, Canesta, RTI, QNX, Novatel, Velodyne (LIDAR). And it is only a matter of time before much more widely recognized names get involved – pick your military contractor and other IT giants such as Samsung, Accenture, IBM (partnership with Magna Electronics), Intel, GE, et al.

Applications we cannot imagine…
Yes, I am excited about being able to fall asleep after leaving New York and waking up in the morning in North Carolina, Maine or Wisconsin. And I’m equally excited to see less stressful commutes, a drop in vehicle fatalities, a reduction in ‘drinking and driving’ incidents, et al.

But I believe the real economic gains of autonomous vehicles will be in areas that are less tenuous than ‘displacing drivers’ on the highways. Freight is always looking for a way to reduce costs. Energy harvesting of remote wind/solar fields is a possibility.

One of DARPA’s sponsors was Caterpillar Construction which ‘now equips some of its bulldozers with a combination of GPS and laser scanners to allow for semi-autonomous earth-moving.” We cannot underestimate the applications for mining and resource extraction.

Why should we care about this now…?
It is time to take the idea more seriously. And have a more optimistic and pragmatic tone when talking with others about this leap forward.

It is almost certain that the technology systems will evolve faster than human organizations. We avoided alternative scenarios developed by futurists in the 1970/80s related to the future of oil supplies (geopolitical issues/peak production) and now we are paying the consequences on catching up.

I believe congestion and the fundamental performance limits of human drivers is next in line- especially given the global sales expected between now and 2050. This is the problem waiting for us to solve in the future. And my belief is that it cannot be solved using human operators. (Yes, yes, of course we’ll drive ourselves on beautiful country roads. But not in urban environments and on congested networks.)

The real question is - how do we have a productive conversation about the role of computer-based transportation systems? And avoid the tendencies to dismiss things are science fiction, too far away for policy makers, or just ‘too scary to think about’.

Video from the qualifying rounds… it’s like watching a baby take its first steps.

…