slides.pug

section
  h1= title
  h2.author= author
  img.logo(src='img/logo.png')
section
  h2 Problems to overcome
  ul
    li.fragment Getting the info early enough
    li.fragment Which direction the cyclist wants to turn
    li.fragment Limit of 100 geofences on Android
    li.fragment Knowing user is a cyclist and not a driver
    li.fragment User experience
    li.fragment Users without Android phones
section
  section
    h2 An app design
    img.featured.less(src='img/app.jpg')
    p (Press space to advance, and type 's' to read notes)
  section
    h2 User registration
    img.mockup(src='img/mockup-signup.jpg')
    aside.notes
      | Users will need to register to use the app. This ensures that we can 
      | identify users who are using the app in a problematic way (e.g. as 
      | drivers) and potentially ban them from the platform.
  section
    h2 Pre-drawn routes
    img.mockup(src='img/mockup-routes.jpg')
    aside.notes
      | In order to understand the direction that users are approaching from
      | and which way they want to turn, the app requires users to follow a
      | pre-entered route in a specific direction. The app should come with
      | several popular commuter routes pre-entered by the council. Although it
      | would be possible to automate the process of working out which direction
      | a person is approaching a signal, there's no way to know which way they
      | want to turn without knowing their route in advance (as we do for buses).
  section
    h2 Initiating a journey
    img.mockup(src='img/mockup-route.jpg')
    aside.notes
      | Once a user has picked a route, they can hit the start button to get 
      | going. The app can confirm their GPS is working and that they are near
      | the start location.
  section
    h2 On approach to a signal
    img.mockup(src='img/mockup-signal.jpg')
    aside.notes
      | When a user approaches a signal, it activates the geofence and sends
      | the appropriate data to the SmartCycling server, which in turn converts
      | it to UTMC and sends it on to the signal priority system. Information we
      | have will include the rider's location, their speed and which direction
      | they intend to turn (based on the current route). The app should look 
      | for every signal on the route, not just the next one, and assume that
      | each signal will only be passed once.
  section
    h2 Feedback in the app
    img.mockup(src='img/mockup-feedback.jpg')
    aside.notes
      | The app can tell people which signals it detected in time to send the
      | prioritization signal. This information should be generated and stored
      | server-side so that it can be used for analytics and fixes to broken
      | routes.
  section
    h2 Auto-ending journey
    img.mockup(src='img/mockup-done.jpg')
    aside.notes
      | If the user passes the last signal, or doesn't pass a signal for a set
      | time (e.g. 1 hour) then it could mark the route as complete. Users
      | could also have a stop button.
  section
    h2 Record mode
    img.mockup(src='img/mockup-record.jpg')
    aside.notes
      | Adding new routes should be possible with a "record mode". This could
      | record which signals a user passes while cycling a new route, and then
      | present them with a list of all signals. Maybe it could assume 
      | "straight ahead" for all of them as auto-detecting the direction would
      | be hard for a small-scale app.
  section
    h2 Manually editing journeys
    img.mockup(src='img/mockup-edit.jpg')
    aside.notes
      | Then users can go in and remove inaccurately detected signals, and 
      | change the direction at signals where they are actually going left or
      | right. There would need to be a UI that is aware of all the signal and
      | movement codes.
section
  h2 Technical details
  ul
    li.fragment Set up Android geofences when journey begins
    li.fragment Send message to SmartCycling server when entering geofence
    li.fragment Message converted to RTIG-UTMC format including "movement"
    li.fragment App needs to be aware of signal IDs and movement IDs
    li.fragment App can use ActivityRecognitionClient to identify drivers looking to abuse priority
section
  h2 Smallest first step
  ul
    li.fragment Identify popular route
    li.fragment Hardcode that route in both directions
    li.fragment Need signal / movement IDs and geofence radius
    li.fragment Build app with just start/stop and route choice
    li.fragment No feedback / editing
section
  h2 Thanks!
  img.featured(src='img/bike.jpg')