A.I. Navigation for (Cycling) Safety and Inclusivity
View: http://dymaptic.com/dy-presentations/cycling-nav-devsummit-2019.html xkcd.comPresented by
Christopher Moravec & Dawn Mott
Based in Portland, Oregon. Creating A Functional Navigation App for Cyclists
Aren't there enough navigation apps out there?
Maybe for cars... and biking for sport... ... but not for city cycling! Why is this important?
systemcomic.comSome Portland Facts:
7.2% of commuters go by bike.
Highest percentage of bike commuters for a large American city
17,000+ workers in Portland choose to cycle
350+ miles of bikeways
77 Neighborhood Greenways miles
188 miles of bike lanes
85 miles of paths
17 miles of bikeways physically separated from automobiles
-Portland Bureau of Transportation, 2016
How do we define our roads? Trail
Protected Bike Lane
Designated Bike Lane Shared Lane
Residential Street Surface Street Highway*
*Our app prohibits highways.
Why a chatbot?
Texting on a bike is silly, right?
The Eventual Goal: Voice-based Interface!Azure Bot Service Chatbot demo
Start getting to know your chat-bot by writing a short scripted conversation. You can give your chat-bot a personality
including a sense of humor.
The Data is Perfect! OpenStreetMap + Portland City Bike Infrastructure Data
Who are these cyclists?Level 1: The Novice or Skittish Cyclist Highest preference towards off street paths and trails. Some protected bike lanes or shared lanes.
Level 2: The Comfortable or leisure Cyclist Protected lanes and off street trails first. Shared lanes and neighboorhood streets second.Level 3: The Daily or Commuter Cyclist Get them there fast!Demo Routing PortionWhat's Next?
Better (more!) data
Voice integration using NLP
AI/ML for preferred route selection
Combining cycling and public transport
Evelation selection options
Thank you. Here's to the future of cycling!