Skip to main content
Andrew Best, Sean Curtis, Dinesh Manocha
Special thanks to David Kasik and Christopher Senesac for models, feedback, and support

Unloading (videos)

  • Initial Global Navigation using Navigation Mesh
    Our initial goal was to unload the plane in a manner intuitively consistent with human unloading. We investigated various forms of global navigation in the aircraft. We decided to use a navigation mesh as our global navigation mechanism. We then began experimenting with social coordination. In this early result, agents are given “right-of-way” when they reach the aisle. The plane unloads in an orderly fashion, however, it does so from back to front.
  • Using Better Right-of-way
    We then updates our “right-of-way” mechanism to reflect the fact that passengers in the front of the aircraft unload before those in the rear. In this simulation, agents are assigned priority based on their proximity to the front of the aircraft. Agents in the aisle are given priority over those in rows.
  • Modelling Behaviors
    Here we’ve modelled some secondary observed behaviors. All agents must stop in the aisle to collect their luggage. The yellow and pink agents will wait until the isles are clear to exit the plane. This gives us more interesting and diverse unloading behavior.
  • Unloading with multiple doors
    Not all unloading situations use a jet-way. In some cases, all the main doors of the aircraft are opened. Here we demonstrate agents choosing the appropriate exit based on their location in the plane.

Loading (videos)

We have also modeled several loading strategies. In these examples, first-class agents are modeled as green discs, and coach agents are modeled as blue disks.

  • Front to back
    In this demonstration, the agents in coach will load front to back.
  • Back to Front
    In this demonstration, coach passengers board back to front.
  • Random
    In this demonstration, coach passengers board in a random order.