Towards a Proximal Resource-based Architecture to Support Augmented Reality Applications. Cynthia Taylor, Joe Pasquale UC San Diego

Transcription

1 Towards a Proximal Resource-based Architecture to Support Augmented Reality Applications Cynthia Taylor, Joe Pasquale UC San Diego 1

2 Background and Motivation Small Devices Big Applications Thin Clients Latency System Architecture Utilities Adaptation of Google Earth Next Steps 2

3 Big Applications Virtual Worlds Maps Augmented Reality Data/Computation Intensive, Context Dependent 3

4 Small Devices Zypad Wearable iphone Netbook Nanotech Contact Lens Display (UW) 4

5 Latency Desktop Thin Client 5

6 Workspace GPS video web user s earpiece PDA Work space game web terminals world 6

7 Second Life Massively Multi-player Online World 7

8 Computer Vision Find and follow objects in video 8

9 Augmented Reality Virtual objects integrated into live video 9

10 Google Earth 3D Ancient Rome Interactively explore ancient Roman buildings 10

11 Application Characteristics Data Intensive Computation Intensive Sensor data Frequent user-interaction 11

12 Background and Motivation System Architecture Terminals World Architecture Utilities Adaptation of Applications Completed Work Next Steps 12

13 Terminals GPS video web user s earpiece PDA Work space game web terminals world A collection of input/output devices and sensors 13

14 World GPS video web user s earpiece PDA Work space game web terminals world Various s scattered over the internet. 14

15 Architecture GPS video low latency high latency web user s earpiece PDA Work space game web terminals world 15

16 The Purpose of the Workspace GPS video web user s earpiece PDA Work space game web terminals world Mediates between world and client, adjusting for performance Quick communication with the client Dependence on physical location runs Add additional functionality to programs Create mash-ups between multiple programs 16

17 Background and Motivation System Architecture Utilities Display Forwarding Sensor Input Rendering Video Augmentation Adaptation of Google Earth Next Steps 17

18 Utilities Based on Survey of Applications Data-Intensive Computation-Intensive Support Real-Time Interaction 18

19 Display Forwarding display user input terminals display forwarder utility workspace world Monitors display updates, forwards them to client. 19

20 Sensor Input GPS accelerometer terminals sensor interpreter workspace world Takes sensor input, translates, forwards. 20

21 Rendering display display forwarder terminals renderer workspace world Receives geometric information, renders it into pixels 21

22 Video Annotation display camera terminals annotator application workspace world combines video with outside information 22

23 Utility Characteristics Offload computation from the client Do not require large alterations to apps 23

24 Background and Motivation System Architecture Utilities Adaptation of Google Earth Next Steps 24

25 Data Model Google Display Application 25

26 Google Earth: Unmodified display GE app Google client 26

27 Google Earth: Workspace video glasses terminals 3D navigator netbook display forwarder GE app sensor interpreter workspace Google world 27

28 28

29 Demo 29

30 Next Steps Focus on I/O issues 30

31 I/O Devices camera microphone mouse accelerometer GPS temperature sensor light sensor RFID barcode reader keyboard biometric sensors touch sensor sound card video card 31

32 I/O Local device application Networked device application 32

33 I/O device driver application device driver network?? application 33

34 Caching Polling Buffering Encrypting Compressing Synchronizing Multiple Datastreams functions Transforming Adding Timestamps Averaging Discarding Non-Recent Updates Predicting Future Updates 34

35 Questions? 35