iccps 2012
play

ICCPS 2012 1 Dionisio de Niz, 1 Lutz Wrage, 2 Nathaniel Storer, 2 - PowerPoint PPT Presentation

On Resource Overbooking in an Unmanned Aerial Vehicle ICCPS 2012 1 Dionisio de Niz, 1 Lutz Wrage, 2 Nathaniel Storer, 2 Anthony Rowe, and 2 Raj Rajkumar 1 Software Engineering Institute 2 Electrical & Computer Engineering Carnegie Mellon


  1. On Resource Overbooking in an Unmanned Aerial Vehicle ICCPS 2012 1 Dionisio de Niz, 1 Lutz Wrage, 2 Nathaniel Storer, 2 Anthony Rowe, and 2 Raj Rajkumar 1 Software Engineering Institute 2 Electrical & Computer Engineering Carnegie Mellon University

  2. Motivation 0 true 0 Collision Avoidance Safety-Critical (Prevent Damage) Flight Control Route Planning Mission-Critical Amount of Intelligence (Provide Value) Takeoff-to-Surveillance Distance Surveillance Coverage 2

  3. Optimization of Mission-Critical Value in Mixed-Criticality Systems Conservative approach over-allocates resources to safety-critical tasks Mission-Critical • Surveillance Coverage Mission-Critical • Takeoff-to-Surveillance Distance Reserve • Amount of Intelligence Information Safety-Critical Reserve Safety-Critical (Based on • Flight Control Worst-Case ) • Collision Avoidance • Route Planning 3

  4. Optimization of Mission-Critical Value in Mixed-Criticality Systems Conservative approach underutilizes resources Mission-Critical • Surveillance Coverage Mission-Critical • Takeoff-to-Surveillance Distance Reserve • Amount of Intelligence Information Underutilized Most of the Time Safety-Critical • Flight Control • Collision Avoidance Safety-Critical Reserve • Route Planning 4

  5. Reclaiming Unused Resources: Overbooking Reclaim unused resources & use them to optimize utility of mission-critical tasks while preserving timing guarantees of safety-critical tasks Mission-Critical • Surveillance Coverage Mission-Critical • Takeoff-to-Surveillance Distance Reserve • Amount of Intelligence Information Used for mission-critical most of the time Safety-Critical • Flight Control • Collision Avoidance Safety-Critical Reserve • Route Planning Overbooking 5

  6. Overloading in Mixed-Criticality Systems Task Period Criticality WCET NCET t 1 Surveillance Cov. 4 Mission 2 2 t 2 Collision Avoid. 8 Safety 5 2.5 t 1 2 2 2.5 2 ½ t 2 4 8 6

  7. Zero-Slack Rate Monotonic Task Period Criticality WCET NCET t 1 Surveillance Cov. 4 Mission 2 2 t 2 Collision Avoid. 8 Safety 5 2.5 t 1 2 1 1 2 2.5 ½ t 2 4 8 Zero-Slack Instant 7

  8. Zero-Slack Rate Monotonic Task Period Criticality WCET NCET t 1 Surveillance Cov. 4 Mission 2 2 t 2 Collision Avoid. 8 Safety 5 2.5 t 1 2 1 1 2 2.5 ½ t 2 4 8 Overbooking 8

  9. Reclaiming Resources in Mixed-Criticality Systems Task Period Criticality WCET NCET Utility t 1 Surveillance Cov. 4 Mission 2 2 {2,2.5} t 2 Collision Avoid. 8 Safety 5 2.5 t 3 Amount of Intelligence 4 Mission 2 2 {2,2.5} Reclaimed t 1 Resources 2 2.5 ½ t 2 t 3 4 8 9

  10. Using Reclaimed Resources to Maximized Utility Task Period Criticality WCET NCET Utility Levels t 1 Surveillance Cov. 4 Mission 2 2 {2,2.5} t 2 Collision Avoid. 8 Safety 5 2.5 t 3 Amount of Intelligence 4 Mission 2 2 {2,2.5} 2.5 Utility 1 1 t 1 1 1 2 1 2 2 2.5 ½ t 2 Resource Total Utility = 2.5 2.5 Utility 2 t 3 1 1 4 8 1 2 Utility Diminishes: Utility ≠ Criticality 10

  11. Using Reclaimed Resources to Maximized Utility Task Period Criticality WCET NCET Utility Levels t 1 Surveillance Cov. 4 Mission 2 2 {2,2.5} t 2 Collision Avoid. 8 Safety 5 2.5 t 3 Amount of Intelligence 4 Mission 2 2 {2,2.5} 2.5 Utility t 1 1 1 2 1 2 2 2.5 ½ t 2 Resource Total Utility = 4 2.5 Utility 2 t 3 1 1 4 8 1 2 ZS-QRAM : More mission-critical utility from same resources 11

  12. 100% 12 ZSQRAM: Period Degradation 10 2 4 8 6

  13. 100% 13 ZSQRAM: Period Degradation 10 2 4 8 6

  14. Two Enforcement Points: Before & After Overload Suspend Period lower utility degradation 14

  15. Two Enforcement Points: Before & After Overload Suspend Period lower utility degradation 15

  16. Metric: Utility Degradation Resilience (UDR) Meet Meet Utility utility deadline? deadline Overload Overload Deadline Misses Deadline Misses Vector Vector 16 + 20 = 36 32 + 10 = 42 Scheduler favors high utility Scheduler does not favor high utility 16

  17. 17 Drone RK (www.drone-rk.org)

  18. 18 AR Drone Hardware

  19. 19 Experiment Setup

  20. Software Structure Safety Critical Low-Level Flight Control Video Aux Sensor Data Sensor Data Task Actuation Task Mission Critical Object Detection Video Streaming Navigation Task Task Task Non-Real-Time … httpd crond ntpd 20

  21. UAV Taskset Parameters C o Task Util 1 Util 2 C T 1 T 2 ZS Actuation 1 1 30 30 Sensor Data 0.1 0.1 65 65 Aux Sensor Data 0.5 0.5 50 50 Navigation 11 11 50 49 Object Detection 7 15 40 100 87 Video Streaming 2 4 10 10 120 40 40 All figures in milliseconds 21

  22. 22 Overloaded Utility Functions Nominal

  23. 23 Demo

  24. 24 Object Detection vs. FPS with RMS

  25. 25 Object Detections vs. FPS with ZS-QRAM

  26. Concluding Remarks CPS requires new scheduling mechanisms that can cope with uncertainty in the environment (variable execution time) Criticality-based overbooking protects safety-critical tasks allowing them to steal cycles from mission-critical ones • But fails to encode diminishing returns within mission-critical tasks ZS-QRAM optimizes mission-critical value • Encoded as concave utility functions • Overbooking within mission-critical tasks Developed a metric that captures the capacity of the scheduler to retain utility in overloads Demonstrated in a surveillance mission on our drone-rk platform 26

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend