GUEST:
No matter how much “rapid transit” we employ, there’s nothing rapid about traffic in big cities. Our choices today are limited — cars, public transit, taxis, and ride-sharing. They all have pros and cons, but not one of them gets people where they need to be quickly.
When will technology finally be used to decrease traffic and not just increase a vehicle’s potential? Perhaps now, thanks to augmented intelligence. Augmented intelligence combines various expert resources with deep algorithms and modeling to enable computer simulations that can help city planners better understand the impact of their decisions related to changing traffic routes, incentivizing driving behavior, developing new public transit systems and routes, and approving new business and housing developments.
Do you have an AI strategy — or hoping to get one? Check out VB Summit on October 23-24 in Berkeley, a high-level, invite-only AI event for business leaders.
To understand how augmented intelligence can help city planners, let’s look at the scope of their dilemma. Public officials in big cities manage multiple transportation types and systems: pedestrians, cars, taxis, bicycles, ride-sharing, buses, rail systems, and more. Managing everything is extremely challenging, leading officials to create specific areas of responsibility, each with its own experts and decision makers. But this siloed approach causes decision makers to focus only on optimizing their own areas. What traffic light and lane changes will reduce gridlock during the afternoon commute? How can we reduce the cost of getting bus riders across town? How can we make metro stations and trains safer?
Working in isolation, however, can cause more problems than it solves. For example, changing a bus route or building a metro stop can cause congestion for cars, just as making traffic light and lane changes can disrupt bus routes and cause new headaches for nearby businesses.
Solving major transportation challenges
Transportation decision makers face three additional major challenges:
- Driver and passenger behavior and expectations continue to evolve. For example, our real-time, on-demand lifestyles make us more impatient, so much so that many drivers insist on using their smartphones while driving, despite the obvious consequences.
- Regional systems are links in global supply chains, and their decisions can cause delays and disruptions in the flow of goods into and out of their areas — with potential national and international consequences.
- Future innovations, such as self-driving cars, drone package delivery and personal and small-group air transport, will continue to cause new disruptions.
This constant and intricate interplay of causes and effects makes transportation a complex system that individual humans cannot fully comprehend. What completely unexpected scenarios — “emergent phenomena” — might arise from the interaction of different systems? What are the cascading effects that might ripple through the entire system from a single decision? A complex system has to be approached holistically, with a much broader and deeper understanding of all the forces at work and all the possible impacts of major decisions.
This is what augmented intelligence is designed to do. Think of it this way — to be an expert in a particular area of the transportation industry, a person must focus predominantly on that area, so the depth of knowledge beyond that area is limited. By extension, a city council making transportation decisions must try to reconcile the often competing and contradictory recommendations of their experts. They must narrow their focus and zero in on what they think is most relevant. In doing so, however, they are ultimately relying simply on their intuition to determine which bits of evidence are most important to their decisions.
Augmented intelligence goes in the opposite direction. It combines many different expert resources — even some that may appear only tangentially related to the problem (such as the trending popularity of a local sports franchise) — into a more comprehensive and sophisticated computer model that actually relies far less on raw data and far more on knowledge to run simulations that explore all the possible causes and effects of a decision. Based on the initial simulation, the computer model is then iteratively refined to allow more and more clarity to emerge by utilizing more knowledge, not by excluding it.
Augmented intelligence software is already being used in the transportation industry. For example, a company focused on smart urban rail design is using an augmented intelligence application to model and simulate railway systems in order to help rail system providers deliver better traffic management and greater energy efficiency. The models and simulations account for the interactions between rolling stock, signaling, infrastructure, and traffic regulation. It answers competitive tenders for rail transportation system design. And it provides guidance related to the ongoing digital revolution in rail systems operation.
Thanks to the holistic view of augmented intelligence, companies using the software are able to reduce headways, increase resilience, better manage disturbance on the network, and improve energy efficiency.
IRT SystemX, a French government research institute, is pursuing its MIC (Modeling-Interoperability-Cooperation) project, focusing on multimodal transportation. Launched in May 2013, the project employs augmented intelligence to optimize multimodal mobility by finding the right balance among transport time, cost, energy consumption, and access to transport across rail, bus, and subway services. It is capable of running simulations that account for the impact of maintenance work, the risk of train delays, passenger numbers, and cumulated passenger delay. This enables transportation decision makers to better balance passenger satisfaction and transport performance. System operators also reap a financial benefit from their better-optimized systems.
What augmented intelligence can and can’t do
Augmented intelligence isn’t artificial intelligence. AI rapidly analyzes tremendous amounts of data about the past to find correlations and make predictions based on those correlations. Augmented intelligence utilizes expert knowledge — not necessarily raw data — to explore a complex system like a transportation network and reveal the emergent phenomena and cascading effects of the interacting systems and activities.
At the same time, augmented intelligence doesn’t predict the future. Instead, it reveals possible futures based on multiple factors and possible decisions, using more sophisticated analytics than has ever before been possible. Still, someone must decide how to use the analysis — what paths to follow. A particular benefit of augmented intelligence is that in showing the cause-and effect-relationships, it enables decision makers to explain their rationale, something that AI typically cannot do.
Ever-growing populations and increasing density are in direct conflict with our desire to move people and products faster. While jammed freeways and gridlocked streets seem to be our fate, augmented intelligence has the potential to help us create smarter, more agile, and free-flowing transportation systems.
Michel Morvan is the cofounder and chief executive officer of Cosmo Tech, a company that helps C-level executives, public leaders, and others make optimal decisions by creating the tools that allow them to account for the complexity that characterizes the world’s most challenging problems.
