Human in the Loop Control.pptx by hcj

VIEWS: 2 PAGES: 10

									Human in the Loop Control
 Victor Shia

 TeleImmersion Lab

 BEARS 2012
                          TeleImmersion Lab




  Tai Chi Teaching         Remote Dancing      Virtual Choreography
  (UCB, Stanford Univ.)      (UCB, UIUC)            (UCB, UIUC)




Portable Teleimmersion      Virtual Geology        Virtual Archeology
         (UCB)               (UCB, UC Davis)    (UCB, UC Merced, UC Davis)
Human-in-the-Loop Control
 Control Theory

 Human input
   Not completely predictable
   In certain scenarios, can be somewhat
    predicted

 Can we develop control algorithms to keep
  systems safe while incorporating nominal
  human input?
    Past Work
 Autonomous cars




 Semi-autonomous cars
    Contributions

     Develop a model for the driver using data learned
      over time

     Apply this model to semi-autonomous cars and
      intervention

                                       behavior
     Illustrate that incorporating human
      improves the performance of a semi-autonomous
      architecture

          R.,
Vasudevan Shia V., et al. Safe Semi-Autonomous Control with Enhanced Driver
Modeling, ACC2012 (accepted)
Experimental Setup
      Models : Results
                          Model 1
                           Normal   Model 2
                                    Distracted



• Model 1: current
  techniques (no driver
  modeling)



• Model 2: driver model
  created from past
  actions and outside
  environment
    Models: Metric


          Accuracy of   Precision of    Recall of      Precision of
          prediction    prediction     intervention   intervention
                                       function       function
Model 1   1             0              1              0.0031
Model 2   0.81          0.82           1              0.68
Conclusion


 Driver modeling can improve the performance
  for the semi-autonomous framework



 Currently implementing model learning and
  prediction system in real time
  Thank You
 475 Hearst Memorial Building

 Demos/Poster
   Gregorij Kurillo
   Aaron Bestick
   Daniel Aranki


                      NSF Grant #0903711 – CiBER -IGERT
 Acknowledgments: NSF,

             :         ,
 CollaboratorsYigi Gao Andrew Gray, Theresa Lin, Prof.
  FrancescoBorrelli

								
To top