Experimental Cooperative Control of Fixed-Wing Unmanned Aerial Vehicles

Size: px

Start display at page:

Download "Experimental Cooperative Control of Fixed-Wing Unmanned Aerial Vehicles"

Oswin Timothy Lambert
6 years ago
Views:

1 Experimental Cooperative Control of Fixed-Wing Unmanned Aerial Vehicles Selcuk Bayraktar, Georgios E. Fainekos, and George J. Pappas GRASP Laboratory Departments of ESE and CIS University of Pennsylvania

2 Outline Design philosophy Test-bed UAV Initiative at Penn Piper J3 Cub Model System architecture (Cloud cap technologies) Hybrid Modeling (of the Piccolo autopilot) Experiments Autonomous Flight Formation Surveillance UAV-UGV cooperation Future directions Temporal logic motion planning

UAVs @ Penn Piper Cub J3 (Scale ¼) Max speed: ~30m/sec

3 Penn Piper Cub J3 (Scale ¼) Max speed: ~30m/sec Nominal speed: ~15m/sec 2.7m Max alt: 5000ft Autonomy: 15-20min

4 Penn Servos controlling the payload GPS receiver LaptopPC Dell X HP fuel engine Deployable pod Pod with IMU 3DM-G and Camera DragonFly Piccolo avionics Antenna

5 Penn Avionics Box Enclosure CloudCap Piccolo Deployable POD (Sensor, Beacon, Landmark, micro-uav) Hi Res Camera & IMU POD Controllable

6 UAV 1 System architecture UAV n Ground Control Airframe Avionics RS 232 CAN Onboard PC Airframe Avionics RS 232 CAN Onboard PC Ground Station RS 232 Operator PC TCP / IP Remote PC

7 Possible Configurations

8 Typical flight plan (mission) Track line segment Service Target Take pictures Orbit Danger Targets Waypoints Turn rate controller

9 Hybrid UAV Control Loop * Piccolo autopilot * Outputs Inputs Mode Switching Hybrid States + Outer Control Loops Continuous Dynamics Airframe Dynamics + Low Level PID Controllers Sensing Models GPS, IMU Pressure Sensors

10 G 32 R 32 G 33 R 33 G 15 R 15 G 11 R 11 Line track (state 1) X& 1 = F1 ( X 1, U1) ) X () t = G( X 1() t ) r r r) ) ω = g x, x, x, θ cmd 1 xtrack > 0 Ÿwpi_cmd Ÿturn_cmd Ÿvel_cmd G 12 R 12 G 13 R 13 G14 R 14 G 16 R 16 ( ) G 34 R 34 G 51 R 51 Circle track (state 3) X& 1 = F1 ( X 1, U 1 ) ) X () t = G ( X 1() t ) r r) r xt = g 2 ( xa, x wp ) r r r) ) ω cmd = g1( xt, x p, xa, θ ) Ÿwpi_cmd Ÿturn_cmd Ÿvel_cmd t p a G52 R 52 Turn Rate track (state 5) X& = F X U ω = u cmd Ÿturn_cmd Ÿwpi_cmd Ÿvel_cmd G54 R 54 G53 R 53 Hybrid model of the Piccolo autopilot ( ) 1 1 1, ω _ cmd 1 G 55 R 55 G 22 R 22 Altitude from WP (state 2) X & = F G 21 R 21 ( X ) 2 2 2,U 2 ŸG 21 Ÿalt_cmd G 41 R 41 Altitude from Cmd. (state 4) X & = F ( X ) 2 2 2,U 2 Ÿalt_cmd Ÿaltfromwp_cmd G42 R 42 G 31 R 31 G 35 R 35

11 Experiments Autonomous Formation Flight (AFF) Simple leader-follower formation Proof of Concept Continuous visual tracking of multiple objects On the ground or in the air Coverage transition from one UAV to another Aerial, mobile sensor network Collective but complementary visual coverage of an area Provide mobile, communication coverage Aerial 2D or 3D mosaics Air-ground integration (ARO MURI) Search in the air, rescue on the ground UAV-UGV cooperation

12 Autonomous Formation Flight (AFF) Fort Benning Aug z (meters) y (meters) y (meters) Starting Point Starting Point Leader x (meters) 100 Leader 0 Follower -100 Follower x (meters) S. Bayraktar, G. E. Fainekos, and G.J. Pappas, Hybrid Modeling and Experimental Cooperative Control of Multiple Unmanned Aerial Vehicles, Technical Report, Department of CIS, University of Pennsylvania, 2004.

13 Autonomous Formation Flight (AFF)

14 Reconnaissance Mission Eye in the sky (150m) Fort Benning Aug Eye in the sky (65m) Aerial mosaic with Rahul Swaminathan

UAV-UGV cooperation Fort Benning, Georgia Nov. 2004 B. Grocholsky, S. Bayraktar, V.

the 9th International Symposium on Experimental Robotics 2004, Singapore, June 2004. B.

15 UAV-UGV cooperation Fort Benning, Georgia Nov B. Grocholsky, S. Bayraktar, V. Kumar, C. J. Taylor, and G. J. Pappas, Synergies in Feature Localization by Air-Ground Robot Teams, Proceedings of the 9th International Symposium on Experimental Robotics 2004, Singapore, June B. Grocholsky, S. Bayraktar, V. Kumar and G. Pappas, UAV and UGV Collaboration for Active Ground Feature Search and Localization, AIAA 3rd Unmanned Unlimited Technical Conference, Workshop and Exhibit, Chicago, Illinois, Sep , 2004.

16 Some Technical Challenges [H. Tanner et. al. 42 * Future nd CDC 2003] Work * Synchronous formations Fixed relative distance or orientation Robust formations critical as UAVs get smaller Cooperative control with sensor constraints Asynchronous cooperation in dynamic environments Include sequencing and concurrency constraints Formal languages for mission/task descriptions Abstractions for controller, sensor primitives Decompose missions into controller primitives Dynamic verifiable composition of components

17 Temporal logic motion planning Building 1 Building 2 Start Ground area UAV go to Building 1 and then to Building 2 and then return to Start and UGV cover Ground Area ( UAV.Building1 ( UAV.Building2 ( UAV.Starting Area))) (UGV.cover(GroundArea) while UGV.avoid(Building1 Building1)) [C. Belta and L. Habets 43 rd CDC 2004] Georgios E. Fainekos, Hadas Kress-Gazit and George J. Pappas, Temporal logic motion planning for mobile robots, Submitted to ICRA'05

18 Temporal logic motion planning Georgios E. Fainekos, Hadas Kress-Gazit and George J. Pappas, Temporal logic motion planning for mobile robots, Submitted to ICRA'05

19 Temporal logic motion planning Georgios E. Fainekos, Hadas Kress-Gazit and George J. Pappas, Temporal logic motion planning for mobile robots, Submitted to ICRA'05

20 Conclusions Development of a fleet of autonomous UAVs Various experiments that demonstrate the operational capabilities of the test-bed Current work in the development of algorithms for the composition of control primitives that satisfy user specifications

21 The END Thank you! Questions?

THE DEVELOPMENT OF A LOW-COST NAVIGATION SYSTEM USING GPS/RDS TECHNOLOGY

ICAS 2 CONGRESS THE DEVELOPMENT OF A LOW-COST NAVIGATION SYSTEM USING /RDS TECHNOLOGY Yung-Ren Lin, Wen-Chi Lu, Ming-Hao Yang and Fei-Bin Hsiao Institute of Aeronautics and Astronautics, National Cheng