Future Mobility – A Perspective
Published by : Industrial Automation
David Bruemmer offers a perspective on smart ecosystem approach for connected vehicle technology to pay dividends in safety and efficiency.
Over one million people die in automobile accidents each year and over 20 million are injured or disabled. Many billions of hours are lost standing still in traffic jams. This does not need to be the case. Imagine a world where vehicles form a virtually interconnected organism, where cars talk vehicle to vehicle, seamlessly accelerating and decelerating together. The challenge is that to achieve this requires a system-level approach including reliable connectivity, accurate positioning and distributed intelligence. GPS fails to meet the requirement especially where we need it most – in city streets. Perhaps nowhere was this clearer than during the rigorous testing performed by the US DOT in New York City where it was not uncommon to find tens of metres of error in urban canyons. The engineers at first thought that there would be significant GPS error only in tunnels and under bridges but in practice most of Manhattan was an “edge case” where GPS failed to work effectively.
The problems with GPS come as no surprise to most technophiles who have long espoused optics-based AI which uses lasers and cameras to build and localise within maps. Despite all the hype, traditional AI has not been all that helpful so far. For years the logic has been that AI needs to progress but the problems are not necessarily with the algorithms, but with the fundamentals behind the approach. Optical sensors like lasers and cameras perform poorly in dust, rain, fog and snow. More importantly, things change from when the map was last made causing confusion. Even if the map remains valid, line of sight sensors will never be able to see around blind corners or even past the next car. Those who love good old fashioned AI say that if we are patient and continue to pour billions into the development of self-driving, then everything will get better, but the limitations of individual intelligence are fundamental. If the Lincoln Tunnel is congested, a single car, not matter how smart, will never be able to think its way through a traffic jam.
It turns out that neither congestion nor accidents are “intelligence” problems. Counterintuitively, it may be that individual intelligence is actually the problem rather than the solution. If you want to know what it would be like to have millions of cars with a really smart brain inside, making individual decisions to maximise efficiency, just look around. It’s what we already have. A number of top thinkers believe that if all cars were imbued with individual autonomy, congestion would actually get worse, rather than better. This isn’t to say that individual intelligence is a bad thing, but it must be complemented by something else if we hope to eliminate congestion and accidents.
Another approach which has progressed along a largely separate trajectory is a connected vehicle strategy, but the challenge here has been that several of the vital requirements are hard to attain. For connected vehicle technology to pay dividends in safety and efficiency requires something even harder than individual intelligence: a smart ecosystem approach. We can define a smart ecosystem as having no more than 10cm position error and no more than 10 milliseconds of connectivity delay. The hard lesson is that getting close to the goal isn’t really helpful. In telecom, 99% reliability made companies billions of dollars. In the Information Age if one call dropped out of 100 no one died. In the Age of Information in Motion, cars are moving at 60mph, 99% reliability means death.
It’s all about the space between cars… We need to think not only about the individual vehicle, but about the collective. Clearly this coordination cannot be done by an individual car, but neither can it be done by off board servers functioning through the cloud. The cloud can’t provide the nuanced timing necessary to coordinate acceleration and deceleration. Moreover, it is vulnerable to hacking and other security risks. If neither centralised controlnor individual intelligence is the answer, what other option do we have?
The answer is to augment individual intelligence (i.e., Tesla autopilot) and centralised intelligence (i.e., Apple maps) with a third and more important capability – swarm intelligence. Paradoxically, swarm intelligence places the burden of control everywhere and nowhere. The distributed approach offers fault tolerance and resilience not possible with any other approach. In the near future, radio chips in each car will range hundreds of times a second to nearest neighbours. To calculate relative position in real time. Simultaneously, cars will talk to radio chips in light posts and other roadside equipment to know exactly where they are on the road. In this sense, we will create a new internet of moving things, where we exchange outdated satellites thousands of miles away in space for a relativistic GPS of things.
Sometimes people ask me why they should care about these abstract concepts, like swarm robotics and peer-to-peer positioning. I tell them it only has to save their lives once to be worth it. We can use swarm intelligence to dramatically reduce greenhouse gas emissions and begin to address fundamental problems like dependence on fossil fuel, traffic congestion, accidents and climate change. As engineers and decision-makers contemplate a move away from the global position paradigm, it may leave them with a strange feeling of vertigo. Like pheromones in an ant colony, our world will fill with digital signposts or “tags.” Some tags will be anchored into the world as infrastructure, built into light bulbs, traffic lights, and signs. If we do this right, traffic lights and stop signs will disappear, replaced by a swarm intelligence that modulates turn-taking with breath-taking efficiency. These infrastructure tags will enhance or simply replace GPS, feeding into existing software applications that expect GPS coordinates. These enhanced apps will guide both humans and robots, creating a common framework for shared understanding and collaboration.
Enhanced positioning enables a shared backbone of safe, coordinated motion. This makes it possible to orchestrate multiple vehicles as a team. In addition to supporting cars, this enables many new forms of autonomous shared mobility. Single-person pods, low-speed electric vehicles, autonomous cars, human driven cars and autonomous ride sharing systems can virtually “snap” together despite their different manufacturers and control systems. Working together these vehicles can smoothly accelerate in unison and perform coordinated, predictive braking when necessary. The idiosyncratic behaviour will evaporate.
Imagine a pod that comes to meet you right at the gate as you disembark from the airplane. It carries you through the airport and merges with another pod that has your luggage before dropping you off directly at the train station or curb side for a car pickup. We can get rid of all the seams – the time sapping transitions between one mode and another. The future won’t be all about self-driving cars, but rather about how to get you from point A to point B as smoothly, safely and efficiently as possible. Whatever comes next, remember that you are focus and that the system should adapt to your needs, not the other way around.
David Bruemmer is CEO and Founder at Adaptive Motion Group. Having worked on intelligent robots for hazardous and critical environments, David is now taking proven technology for positioning and autonomy into transportation and a variety of industrial smart