IBM Research Initiative Developing Adaptive Systems to Provide Personalized Travel Routes to Avoid Gridlock
27 February 2010
IBM has launched a new research initiative to build personalized travel routes for commuters to avoid traffic gridlock. IBM researchers are using advanced analytics to develop adaptive traffic systems that will intuitively learn traveler patterns and behavior to provide more dynamic travel safety and route information to travelers than is available today.
According to the Texas Transportation Institute, as cited by IBM, US traffic congestion burns enough fuel every year to fill 58 supertankers and wastes enough time to consume 105 million weeks of vacation.
New models will predict the outcomes of varying transportation routes to provide a personalized recommendation that get commuters where they need to go in the fastest time. This project intends to provide information that goes well beyond traditional traffic reports, after-the fact devices that only indicate where you are already located in a traffic jam, and web-based applications that give estimated travel time in traffic.
Using new mathematical models and IBM’s predictive analytics technologies, the researchers will analyze and combine multiple possible scenarios that can affect commuters to deliver the best routes for daily travel, including many factors, such as traffic accidents; commuter’s location; current and planned road construction; most traveled days of the week; expected work start times; local events that may impact traffic; alternate options of transportation such as rail or ferries; parking availability; and weather.
Working with state and local transportation agencies, IBM plans to launch pilot projects for select sets of commuters to analyze, test and refine the new systems. IBM plans to provide program participants with the personalized commuting information via the web, through mobile voice interaction, combined with advanced mapping applications on mobile devices.
For example, combining predictive analytics with real-time information about current travel congestion from sensors and other data, the system could recommend better ways to get to a destination, such as how to get to a nearby mass transit hub, whether the train is predicted to be on time, and whether parking is predicted to be available at the train station. New systems can learn from regular travel patterns where you are likely to go and then integrate all available data and prediction models to pinpoint the best route.
Insight from IBM’s analytics and pilot programs will help transportation agencies better understand and manage traffic, increasing safety on our roads and encouraging the use of efficient public transportation which will help reduce a commuter’s overall carbon output.
The data exists to give commuters and transportation agencies a better way to manage traffic but today it’s not connected. IBM has the ability correlate all of this information to better predict demand, optimize capacity help improve traveler and highway safety as well as reduce our impact on the environment.
—Gerry Mooney, General Manager, Public Sector, IBM
Additionally, IBM is launching a new global virtual Travel and Transportation Center of Competency which will provide new solutions and deep industry expertise for air, rail, truck, and sea transportation.
All well and good. And interesting use of predictive systems data. But will this system get me to the office if I have a commute on the LA 405 freeway? There ARE no alternative routes that don't simply add to the time.
On the other hand, should I be driving a Leaf, Tesla or Volt, I consume little energy during idle times. Another reason for fleet transition to EVs.
I am all for these kind of routing solutions, but as urban centers continue to grow and with finite transportation corridors - how will this really help? One solution may be to dedicate two or even three lanes to priority vehicles - HOV and or low fossil fuel vehicles.
LA seems a good place to pilot an automated high speed commuter lane. This requires off the shelf hardware to sequence vehicles on and off a dedicated lane that automates the drive. The driver manually accesses the lane and once on it - all steering, separation, and braking functions are automated. Once the vehicle reaches its destination off ramp - manual control is returned to the driver. Simply a souped up cruise control technology.
Vehicles capable of this transit mode would need to conform to a standard designed to provide automated vehicle control and redundant safety systems. DOT, DOE, EPA and Commerce should all contribute funds to finance this program. Reconfiguring high density traffic lanes to automated lanes will speed the adoption of EV and hybrid vehicles. Who wouldn't want a twenty minute commute in place of an hour and fifteen?
Posted by: sulleny | 01 March 2010 at 10:40 PM