Software Enabled Control for Intelligent Uninhabited Air Vehicles UAVs

1
Software Enabled Controlfor Intelligent
Uninhabited Air Vehicles (UAVs)
Contract Number F33615-98-C-1341 Award End
Date 3Q-FY04

• Principal Investigators
• Daniel Schrage (AE) George Vachtsevanos (ECE)
• School of Aerospace Engineering School of
  Electrical and Computer Engineering
• Georgia Tech Georgia Tech
• Atlanta, GA 30332 Atlanta, GA 30332
• daniel.schrage_at_ae.gatech.edu george.vachtsevanos_at_e
  ce.gatech.edu
• [phone] [phone]
• Co-PIs and Key Personnel
• Bonnie Heck (ECE), Eric Johnson (AE), J.V.R.
  Prasad (AE), Linda Wills (ECE)

controls.ae.gatech.edu/projects/sec
2
Problem Description Objectives

• Develop software-enabled control methods for
  complex dynamic systems with application focus on
  intelligent UAVs
• Support-the-development and implement a
  plug-and-play, real-time software architecture
  for SEC integration (OCP-based)
• VTOL UAV hardware-in-the-loop simulation and
  flight test
• University experiment lead for SEC demonstrations

3
Project Tasks

• Task I Hybrid controls API development
• Develop and use transition management
• Adaptive control/online customization support
• Limit detection and avoidance support
• Task II Flight testing of OCP integrated control
  algorithms for achieving high confidence
  performance in unmanned aerial vehicles
• Task III Technology transfer / SEC university
  experiment lead

4
Overview of Approach
Control Algorithms
OCP and HybridControls API
UAV Flight Test Demonstration Using OCP
Modeling and Simulation - SILS HILS
UAV PlatformIntegration
5
Status in May 2002
Released June 2002
Navigation System
Developed Tested SITL
Low Level Controller
HITL
Flight Test
Flight Computer
Instrumentation
Mid-Level Algorithms running in simulation (e.g.,
Matlab)
Limit Avoidance
FDI
Running on OCP 2.0, NT
Mode Transition
Running on OCP 2.0, NT
Transition Management (Hybrid Controls API)
6
November 2002
Previous accomplishment New accomplishment In
progress
Released June 2002
Navigation System
Developed Tested SITL
Low Level Controller
HITL
Flight Test
Flight Computer
Instrumentation
Release Reference Architecture
Data Link Interface
SITL OCP on NT
SITL QNX
HITL
Flight Test
Trajectory Generation
Limit Avoidance
FDI
2nd Flight Computer
SITL OCP on NT
Mode Transition
OCP, NT
Transition Management (Hybrid Controls API)
OCP, NT
Release in OCP Build 2.2
7
Subcontractors and Collaborations

• Collaboration with Boeing on OCP
• Requirements
• Controls API
• OCP releases
• Support
• Flight testing
• Hybrid Controls API

8
Subcontractors and Collaborations (cont.)

• Oregon Graduate Institute (OGI)
• Principal Contact Dr. Richard Kieburtz
• Task Autonomous GTMax avionics and vehicle
  control under high-performance maneuvers
  environment-informed sensing and control in OCP
• Honeywell (High Confidence Project)
• Principal Contact Dr. Tariq Samad
• Task Demonstrate statistical verification of
  computational or performance aspects of on-line
  GTMax functions embedded in OCP

9
Subcontractors and Collaborations (cont.)

• Vanderbilt University (VU/ISIS)
• Principal Contact Dr. Gabor Karsai
• Task Evaluate fault-adaptive control technology
  elements in GTMax use Generic Modeling
  Environment (GME) to generate C code for an OCP
  controller
• Draper Laboratory
• Principal Contact Dr. Felsa Satlow
• Task Demonstrate autonomous aggressive maneuver
  trajectory generation on GTMax integrate
  aggressive flight control technology into GTMax
  demo platform.

10
Subcontractors and Collaborations (cont.)

• MIT
• Principal Contact Dr. Eric Feron
• Task Demonstrate on GTMax high-confidence
  properties of multimode control using hybrid
  input-output automata analysis techniques
• University of Minnesota
• Principal Contact Dr. Gary Balas
• Task Anytime Control Algorithms - mode switching
  framework and reconfiguration strategy for
  autonomous UAVs

11
Subcontractors and Collaborations (cont.)

• Scientific Systems Company, Inc.
• Principal Contact Dr. Jovan Boskovic
• Task Demonstrate through GTMax flight tests mode
  switching model predictive control, FDIR,
  autonomous trajectory generation, path planning
  and decision making algorithms
• Impact Technologies, LLC (DARPA SBIR Project)
• Principal Contact Dr. Michael Roemer
• Task Demonstration of automated contingency
  management software on GTMax

12
SEC University Experiments Plan A Grand
Challenge for DARPA Information Exploitation
Office (IXO)
GT-HW-MIT VVA and High Confidence Systems
GT-VU-SSC Mode Transitioning Fault Tolerant
Control
GT-OGI-Draper-SSC-UMN Trajectory Generation
Extreme Maneuvers

Time Critical lt15min
Target Engagement Kill
Identify Structure Portals
Boeing-GT-VU-UCB-HW OCP Enhancements Hybrid
Controls API
13
Status Hybrid Controls API

• Transition Manager

14
OCP Transition Management for Hybrid Controls
15
Status

• Refining documentation and basic examples
• Target release Build 2.2
• Pre-released to Vanderbilt
• Planned experiments with
• Adaptive mode transition controller
• Multiple limit avoidance w/ changing relative
  importance
• Fault tolerant control
• Guler, M., Clements, S., Kejriwal, N., Wills, L.,
  Heck, B., and Vachtsevanos, G., "Rapid
  Prototyping of Transition Management Code for
  Reconfigurable Control Systems," Proc. IEEE Int.
  Wksp on Rapid Systems Prototyping (RSP),
  Darmstadt, Germany, pp. 76-83, July 2002.
• Guler, M., Clements, S., Wills, L., Heck, B., and
  Vachtsevanos, G., "Generic Transition Management
  for Reconfigurable Hybrid Control Systems," to
  appear in IEEE Control Systems Magazine, 2003.

16
Status Control Algorithms

• Adaptive Mode Transition
• Fault Tolerant Control
• Limit Detection and Avoidance

17
Integration of Mission Planning and Trajectory
Generation Routines with GTMax Software
18
SITL And Flight Test Results
Flight Test
SITL simulation
19
Integration of the Adaptive Mode Transition
Control Architecture / Simulation Results
Desired and Actual 3D Trajectory
Commanded and Actual Position and Velocity
Turbulence
Mode Transition For Moving Obstacle Avoidance
Target
1
MovingObstacle
2
20
Integration of the FDI / Fault Tolerant Control
Architecture with GTMax Software
OCP on secondary computer
Fault Detection Identification
High-level Controls
Fault information
Fault TolerantControl Reconfiguration
Mid-level Controls
Process 1
InterconnectionReconfiguration
SetpointReconfiguration
LL ControllerReconfiguration
Reconfiguration Commands
Datalink
UAV state
21
Bob-Down Maneuver withStuck Collective Fault
FDI Results from SITL
22
Envelope Protection for UAVs
Problem Extreme performance maneuver commands
might cause control inputs that could result in
limit violations
Solution Use a system that can detect
approaching limits and prevent the vehicle
from exceeding them

• Related Publications  
•   I. Yavrucuk, J.V.R. Prasad, Adaptive Limit
  Margin Prediction and Control Cueing for Carefree
  Maneuvering of VTOL Aircraft, AHS Flight
  Controls and Crew System Design Technical
  Specialists' Meeting, Philadelphia, PA, Oct.
  2002.
• I. Yavrucuk,  J.V.R. Prasad,  A.J. Calise, S.
  Unnikrishnan,  Adaptive Limit Control Margin
  Prediction and Avoidance, 58th AHS Annual Forum,
  June 2002.
• J.V.R. Prasad, I.Yavrucuk, S. Unnikrihnan
  Adaptive Limit Prediction and Avoidance for
  Rotorcraft,28th European Rotorcraft Forum,
  Bristol, UK, Sept. 2002.

23
Limit Avoidance System Architecture
.

Approximate Model
Helicopter Controller
?
-



uc

e1
K
?
Adaptive Neural Network
uc
Model Generation Loop
Dynamic Trim
yplim
Allowable Command Margin
24
R-Max Flight Test Results(Rotor Stall Limit
Avoidance)
Hover-to-hover rapid transition maneuver
Predicted command margin violation precedes
actual limit violation
Limit Avoidance ON
Limit Avoidance OFF
25
Status Flight Tests and Demonstrations

• GTMax Research UAV Development
• Flight Testing Results

26
Georgia Tech R-Max GTMax

• Yamaha R-Max,
• 66kg
• 3m Rotor
• Flights Began March 2002

• Instrumented as a Research VTOL UAV
• OCP
• Modular Avionics Hardware

• Related Publications  
• Johnson, E. and Kannan, S., Adaptive Flight
  Control for an Autonomous Unmanned Helicopter,
  Proceedings of the AIAA Guidance, Navigation, and
  Control Conference, 2002.
• Dittrich, J. and Johnson, E., Multi-Sensor
  Navigation System for an Autonomous Helicopter,
  Proceedings of the 21st Digital Avionics Systems
  Conference, 2002.
• Kannan, S. and Johnson, E., Adaptive
  Trajectory-based Flight Control for Autonomous
  Helicopters, Proceedings of the 21st Digital
  Avionics Systems Conference, 2002. (Awarded Best
  Student Paper)
• Christophersen, H., Hart, M., Dhingra, M., Guily,
  R. and Johnson, E., Development of an Autonomous
  Aerial Reconnaissance System at Georgia Tech,
  Proceedings of the Association for Unmanned
  Vehicle Systems International Unmanned Systems
  Symposium Exhibition, 2002.
• Johnson, E. and Mishra, S., Flight Simulation
  for the Development of an Experimental UAV,
  Proceedings of the AIAA Modeling and Simulation
  Technologies Conference, 2002.

27
Since Last Time

• Since the last PI meeting (June-Oct) the GTMax
  has had
• 12 flight test days (about every other week)
• Approximately 53 flights, 14 hours in the air
• Approximately 37 flights where the autopilot was
  utilized
• 70 segments of recorded data archived
• 1 primary flight computer operating system change

28
Aerial Robotics Competition
July 31 August 1, 2002
29
Goal Level 2 Mission
Find Building and Map Entry Points using Onboard
Systems
Launch Area
Image Receiver (Just to see if its working)
Sign on Correct Building
gt1m
30
Flight Configuration (Aerial Robotics)
Actuator Raw Data
Sensor Raw Data
Sensor Drivers
Actuator Driver
Sensor Data
State Estimate
Flight Controller
Navigation Filter
Control
Command Vector
Tracker
Flight Computer
Second Computer
Target Parameters State Estimate Image
Processing Results
Onboard Camera
Desktop Computer
Ground Control Station
31
GTMax Finds the Three Buildings
32
Aerial Robotics Results

• Mapped all three buildings on three attempts
• Was able to complete window mapping of all three
  buildings once (due to GPS failures during two)
• On this run
• Correct building identified
• Window found, unfortunately on a neighboring
  building only 12 ft away!
• Based on this performance (autonomous flight,
  mapping of buildings and windows) Georgia Tech
  team wins!

33
Baseline Control/Navigation System Flight Testing

• Envelope expanded 50 to 80 ft/sec speed
• Have experienced up to 40 knot gusts (estimated)
• First automatic takeoff and landing
• First automatic aggressive maneuvers

34
Flight Configuration (October Tests)
Actuator Raw Data
Sensor Raw Data
Sensor Drivers
Actuator Driver
Sensor Data
State Estimate
Flight Controller
Navigation Filter
Control
Command Vector
Primary Flight Computer
OCP Mid-Level ControlComponents
Navigation DataTrajectory Commands
Second Computer
Desktop Computer
Ground Control Station
35
Flight Computer Interface
Signal LimitInfo Position(x,y,z) Velocity(x,y,z)
Acceleration(x,y,z) Time Delf, Delm(x,y,z) Rpm
Limit Avoidance Component
50 Hz
Low Level Flight Controller, Navigator
Datalink Component
LimitInfo
LimitDecl
Signal LimitDeclaration AccelerationLimit(x,y,z)
Ethernet UDP Comm
UAVState
50 Hz
Mission Trajectory Planning Component
SetPoint
25 Hz
Signals UAVState SetPoint Position(x,y,z) Veloc
ity(x,y,z) Phi, Theta, Psi p, q, r
Primary Flight Computer
Secondary Flight Computer
36
Trajectory Planner (Start)
37
Limited Detection and Avoidance
38
First Automatic Takeoff
39
First Automatic Landing
40
Automatic Aggressive Maneuver
180 Degree Velocity Change in a congested
environment
Keep nose aligned with velocity (zero sideslip)
throughout
Start and Finish at 30 Knots
41
1800 Velocity Change
42
Project Status Schedule
43
Project Plans (6 months)

• Perform tests for real-time performance of
  transition manager, integrate existing control
  laws into OCP
• Flight test mid-level control algorithms using
  the transition manager for the hybrid controls
  API
• Adaptive mode transition controller
• Fault tolerant controller
• Automatic limit detection and avoidance
• Coordinate mid-term and final experiment
  activities with SEC collaborators

44
Program Issues

• Coordination of university-led exams/experiments
• Detailed plans (version 2.0)
• Interface Requirements
• Schedule
• Open control platform development to support
  university-led mid-term and final
  exam/experiments