1 / 38

Optimal Policies in Complex Large-scale UAS Traffic Management

Explore the futuristic vision of UAS traffic management, dynamic UAV flight path planning in urban environments, using BRECCIA agents for optimal policies in complex scenarios. The system includes components like BDI Engine, Hadoop GeoWave Connector, and specialized functions for real-time decision-making. Discover the state transition, reward function, and value iteration algorithm. Advances in mapping, delivery, inspections, and urban air mobility are discussed. Learn more about UAS fleet and large-scale simulations at ICPS 2019.

retar
Download Presentation

Optimal Policies in Complex Large-scale UAS Traffic Management

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Optimal Policies in Complex Large-scale UASTraffic Management David Sacharny & Tom Henderson ICPS Taipei, Taiwan 7 May 2019 ICPS 2019

  2. Futuristic Vision (Slide from Jared Esselman; UDOT) ICPS 2019

  3. Commercial Use Cases • 3D Mapping, Video Collection • Delivery (Amazon, etc.) • Inspections • Data (Re)Transmission • Air Taxis • Investment 2017: $506M • 1000’s of flights per day ICPS 2019

  4. Drone HW Investment ($B) ICPS 2019

  5. Utah Urban Air Mobility Idea (Slide from Jared Esselman; UDOT) ICPS 2019

  6. UDOT UAM (cont’d) (Slide from Jared Esselman; UDOT) ICPS 2019

  7. UDOT UAM (cont’d) Proposal: Airwaysabove roadways. ICPS 2019

  8. UAM: Need to Plan Flights ICPS 2019

  9. Dynamic UAV Flight Path Planningin Urban Environments

  10. Geospatial Intelligence:BRECCIA ICPS 2019

  11. BRECCIA ICPS 2019

  12. URBAN Implementation BRECCIA Agent • The BRECCIA Agent represents the core abstraction for all agents in the system. • Agents are distributed across specialized machines such as UAVs, mobile laptops, or high performance computers. • The inherited components of each BRECCIA agent enable an overall system that is dynamic and data-driven. BDI Engine :Mission Planner Uncertainty Reduct. Goal DB P.L. Logic Module Hadoop GeoWave Connector RRT* Planner Accumulo Specialized Functions GeoServer :Weather Monitor :UAV Manager :User Example Instantiations of the BRECCIA Agent ICPS 2019

  13. Which Path To Take? What about Wind? What about Rain? ICPS 2019

  14. Airway Corridors E.g., over Salt Lake City

  15. Airspace Volumes Z Y -X X -Y -Z (b) Action Directions (a) Airspace Volumes

  16. Learning Optimal Action Policy 4x4 Grid ICPS 2019

  17. Bellman Equations where: U(s) is the utility of state s a is an action A(s) is the set of actions in state s R(s) is the reward for state s is a horizon coefficient ICPS 2019

  18. State Representation state space: * 3 integer grid coordinates * 3 wind vector values (x,y,z) * 1 precipitation value * 1 temperature value ICPS 2019

  19. State Representation:Reduced state space: * G = {1,2,3,4}: grid indexes * W = {0,1}: no wind; wind * P = {0,1}: no rain; rain * T = {0,1,2}: cold, normal, hot (temp) ICPS 2019

  20. Actions A = {X, -X, Y, -Y, Z, -Z} * move in one of the coordinate directions ICPS 2019

  21. Probabilistic State Transition ICPS 2019

  22. Reward Function ICPS 2019

  23. Value Iteration Algorithm From: Russell & Norvig ICPS 2019

  24. Experiments • Start Location: 1,1,1 (index 1) • Goal Location: 4,4,4 (index 64) • Blocked Cell: 4,4,3 (index 60) • Can’t exit 4x4x4 • Preference for horizontal motion ICPS 2019

  25. State Utilities and Path ICPS 2019

  26. Convergence for Utilities Grid cell [3,4,4] [3,2,4] [2,1,1] ICPS 2019

  27. Optimal Policies X: RIGHT Z: UP Y: BACK ICPS 2019

  28. Optimal Policies ICPS 2019

  29. Cell Travel Density ICPS 2019

  30. Policies with Wind in Y No action in Y axis! No Wind Strong Wind in Y Direction ICPS 2019

  31. Current Work: Get Data! ICPS 2019

  32. Current Work: Testing!Deseret UAS ICPS 2019

  33. Conclusions • Developed effective and efficient optimal policy method • Converted core BRECCIA system to work for UAS Traffic Management • allows communicating, autonomous agents • Cloud computing ICPS 2019

  34. Current Development ICPS 2019

  35. Large-scale Simulation • http://www.cs.utah.edu/~cem/uav/

  36. Questions? ICPS 2019

  37. UTAH UAV Fleet ICPS 2019

  38. Acknowledgment This material is based upon work supported by the Air Force Office of Scientific Research under award number FA9550-17-1-0077 (DDDAS-based Geospatial Intelligence) ICPS 2019

More Related