Maturing UAV Capabilities

- Stepping from Technology Demonstrators to Mission-Specific Systems

KC Wong, DM Newman, PW Gibbens, DJ Auld,
S Wishart, H Stone, JAG Randle, KS Choong, DP Boyle, PW Blythe
Department of Aeronautical Engineering, Building J07,
University of Sydney,
NSW 2006, AUSTRALIA.
Tel: 61-2-9351 2338 Fax: 61-2-9351 4841
<CLICK here for WEB Powerpoint Presentation>

SUMMARY

The Department of Aeronautical Engineering at Sydney University has been conducting research and development on Remotely Piloted Vehicles (RPVs) or Unmanned Aerial Vehicles (UAVs) since 1988 when design was initiated on RPV airframes, and a large scaled model of the Citabria was built as a technology demonstrator. Research work to date has produced promising results towards the development of fully autonomous capabilities for unmanned aircraft, and has brought the core autonomous flight control system to an advanced stage of development. An aircraft currently being operated is the RPV named Ariel. Being developed primarily to provide a flight research platform in support of the department's various research activities, Ariel is also used to enhance skills in airframe design and fabrication, instrumentation, flight control systems, and operational aspects of UAVs. It forms the basis of a technology demonstrator for many aspects of a generic UAV system development and operation. Research has shown strong indications that UAV technologies are mature enough for mission-specific systems to be developed.

Current RPV/UAV related research activities at Sydney University include the following:

1 BACKGROUND

As a preliminary RPV technology demonstrator, a one-quarter scale model of a typical light aircraft was built in 1988 to examine its spinning characteristics with a leading-edge device that was developed in this department. However, off-the-shelf instrumentation and telemetry equipment to meet low-cost requirements were not then available. Low cost sensors and associated telemetry equipment were thus developed and tested. Results from the flight tests were encouraging, showing that high-value research, using small unmanned aircraft, is achievable on a minimal budget.

As development progressed, it became apparent that a larger, heavier, more modular (and thus easily modifiable) RPV would have applications not only to the department's immediate research aims, but also to other projects. Since presenting this work at the Australian Aeronautical Conference in 1989, several research groups have expressed interest in this work.

2 MISSION SPECIFIC REQUIREMENTS

2.1 Introduction

With UAV technologies and capabilities within the department maturing, attention is turning to the tailoring of these technologies to the possible missions in which autonomous UAV's will be required in the near future. The study of suitable airframes and flight control systems for different applications can best be achieved by a group dedicated to RPV/UAV research. Without the constraints of fixed marketing objectives, the university or dedicated research centre provides an excellent environment for objectively testing maturing technologies. Currently there are about ten academic staff and postgraduate researchers actively working on various aspects of UAVs, forming one of the largest UAV research teams in the country. Research has shown strong indications that UAV technologies are mature enough for mission-specific systems to be developed.

In addition to advanced design and construction, departmental UAV activities are progressively incorporating autonomous capabilities in mission planning, trajectory optimisation and flight control. These capabilities will greatly increase the scope of mission profiles, while providing a reliable automatic pilot capability that would not require a prepared landing site, expert pilot, special ground equipment or ideal weather conditions. Potential applications for such a system include coast watch, geological surveying, search and rescue operations, reconnaissance, remote delivery of urgent equipment/material, resource assessment, environmental monitoring, and any other mission that may benefit from autonomous operation.

Applications fall within a number of mission areas, each of which is characterised by particular UAV system requirements.

2.2 Flight Research Platform

The UAV systems under development will be used as test-beds for experiments researching into aerodynamics, stability and control. These experiments will require precise repeatable manoeuvres to be performed, hence:

2.3 Surveillance, Reconnaissance and Atmospheric Monitoring

These missions require long endurance and high reliability flight. Due to the ranges covered, the flight control must have:

2.4 Geological Survey, Resource Assessment and Environmental Monitoring 

In these roles the following are required of the UAV system:

2.5 Conclusion

Many of these requirements have received extensive research attention in the recent past (see publications). Current research and development within the Department of Aeronautical Engineering at Sydney University is focussing on detailed configuration design and control system requirements which will make above missions completely achievable by the UAV systems under development within the department.

3 CURRENT RESEARCH ACTIVITIES

3.1 UAV Airframes

A one-quarter scale model (13.5 kg, 2.68 m wingspan)of the Bellanca Citabria has been used successfully as an RPV technology demonstrator (KCEXP-3), initiating the current research on unmanned aircraft.

Another small RPV of 7 kg AUW and 2.38 m wingspan has been used in the investigations of using wing leading-edge devices to prevent stall in light aircraft, and is being used to investigate parachute recovery techniques of ultralight aircraft.

Preliminary design studies have been done on a Solar-powered High Altitude Long Endurance UAV (Solar HALE). A half-scale unpowered model has been built and test-flown.

Since work commenced on RPV/UAV systems in 1989, it quickly became apparent that a modular RPV would be of value not only to the department's immediate research aims, but also to many other applications. The RPV Ariel was thus developed. It has a wingspan of 3.02 metres and a maximum takeoff weight of 36.5 kg, and is being powered by a 4.1 kW engine. Ariel is currently remotely piloted, but an autonomous flight controller is in an advanced stage of development.

3.2 System Hardware and Software Development

3.2.1 Flight System Hardware

The RPV Ariel currently has an onboard sensor package which includes:

These sensors interface with a Pentium 90 computer. A spread spectrum radio modem onboard is used to send the flight data to the ground station where it is recorded and displayed in real time. The data is also recorded onboard the aircraft in a RAM disk, which will be replaced in the future with a solid state flash disk. The computer will be used to interface the sensors to the flight control software and telemetry.

3.2.2 GPS Receiver Interface Software

The receiver software for the GPS has been written in three modules: the GPS Binary; the Data I/O; and the User Interface. This code modularity allows different hardware/software configurations to be accommodated with a minimal amount of re-coding. The current version of the software running onboard the RPV combines the GPS Binary and Data I/O modules to form a software unit which can control multiple GPS receivers and outputs the navigation data, after simple pre-processing, in a predefined format for use by guidance and control software. Another module of the GPS software used on the ground receives the raw GPS data from the RPV, via a radio modem link, and combines this data with a ground based GPS receiver. A "Relative Position Difference" solution is calculated to correct for atmospheric errors. This corrected data is graphically displayed as instantaneous total velocity, vertical velocity and altitude bar graphs, a trajectory trace map, and total velocity and altitude time history trace. Other information displayed as text is current position in latitude/longitude coordinates and the number of satellites visible to the ground and airborne GPS receivers.

3.2.3 GPS Attitude Determination

In order to achieve the position and altitude accuracy required for survey work, high precision measurement equipment is needed, both to correlate environmental measurements with position, to ensure complete and even coverage of the survey area, and to provide accurate positional reference for the control systems. GPS is being used to give periodic (10 Hz) position and attitude measurement updates to augment a strapdown Inertial Navigation System consisting of accelerometers and rotation rate sensors. The INS is employed to give short term manoeuvre accuracy, while the GPS provides long term positional correction. Multiple GPS receivers are being used to generate measurements of the aircraft orientation. These are used in the state estimator to provide accurate state estimates.

3.3 UAV Flight Simulation

The RPV Ariel continued a development process built on previous experience gained with a succession of smaller vehicles. As part of the process, analytic studies and wind tunnel modelling were carried out. The results of these studies were incorporated in a personal computer based, six degree of freedom, real-time, non-linear simulation of the RPV's flight dynamic and performance characteristics. The simulation was to permit analytic examination of proposed RPV tasks, provide a database for development of control laws and command structures, generate "flight data" for use in various flight dynamics analytic processes, and generally serve as a tool in departmental flight dynamics teaching and research.

An early version of the simulation was supplied under contract to DSTO for use in operational studies. Since then, the simulation has been refined and expanded. It now includes a full low altitude turbulence and wind shear model based on MIL-F-8785 and the von Karman turbulence spectrum, and undercarriage and ground plane models. These have been checked against flight data for several takeoff and landing cycles. Engine/propeller effects modelling has also been improved, based on both experimental and computational data.

3.4 Guidance and Control System Research and Development

3.4.1 Introduction

Control system development is one of the most critical aspects of UAV development. To achieve the stringent control precision, performance and guidance requirements posed by some of the missions discussed above, a number of control system components are crucial.

3.4.2 State Estimation and Multi-Rate Sensor Data Fusion 

Due to the presence of noise and measurement systematic errors in the measurement signals, a state estimator has been developed based on a multi-rate Extended Kalman Filter algorithm to provide optimal filtered state estimates to the state feedback control system. This system will provide accurate state estimates by optimally removing noise from the inertial measurement signals, by identifying instrument systematic errors where necessary in order to minimise state estimate drift, and by correcting for positional drift.

3.4.3 Guidance and Control

Numerous control strategies have been developed for the task of UAV manoeuvre and trajectory control. The type of control principle employed may be dependent upon the flight phase, the control performance requirements and the accuracy with which the stability and control characteristics of the aircraft are known. Classical, robust, nonlinear and neural network control principles have been investigated and implemented for Ariel

3.4.4 Trajectory Optimisation

Significant research has led to the development of optimal algorithms for flight trajectory determination given a set of objectives/way-points and an environment through which the aircraft must navigate, which may include obstacles such as mountains and buildings. The algorithm is amenable to real-time implementation and will become a part of the autonomous control system to enable on-board trajectory redesign should the aircraft encounter unexpected obstacles.

3.4.5 System Identification

A research program was undertaken into the use of neural networks for system identification of aircraft aerodynamics, which was targeted towards developing online, adaptive methods for use in UAVs. The work covered the estimation of aerodynamic coefficients for use in fault detection and reconfiguration, and as part of a robust nonlinear control system. In addition, it can also be used offline for model verification from flight test data. The system works by using a supervised training algorithm to train a recurrent neural network in real time, and then using a matrix psuedoinverse to collect the linearised derivatives from the plant estimator.

4 SPECIFIC UAV DEVELOPMENT PROJECTS

4.1 RPV Ariel


First flown in 1995, the flight testing of the department's research Remotely Piloted Vehicle, RPV Ariel continued through 1996. The onboard Global Positioning System (GPS) receiver was used to gather data on the performance of the aircraft. The Ground Control Computer (GCC) software provides a real-time graphical representation of the aircraft flight path, with graphical and numerical data for velocity, altitude and rate of climb. A sample screen is shown below.

4.2 Solar Powered High Altitude Long Endurance (Solar HALE) UAV

The Solar HALE was initiated to study the feasibility of using solar-power in a small UAV capable of carrying a small meteorological package up to 30 km altitudes and capable of being airborne for extended periods. A preliminary study resulted in a 15 kg UAV with a swept flying-wing configuration of 10.95 m wingspan (Figure 6). Conventional technology propellers would, however, restrict its operating altitude to less than 14 km. An unpowered half-scale model has been built and flown, in conjunction with wind-tunnel model tests, to investigate control and stability characteristics.

4.3 Optimisation of a Tail-Sitter UAV

Tail-sitter UAVs represent one solution to the problem of combining helicopter low-speed characteristics with fixed wing high-speed flight. As such, they offer the possibility of obtaining the operational flexibility of the helicopter without its performance limitations in terms of range, endurance and maximum speed. Current research work in the department involves looking at the configuration optimisation of a particular tail-sitter vehicle for a given mission requirement.

The optimisation process uses a parametric model of the vehicle combined with propeller, aerodynamic, structural, engine and hover-control analysis routines. For any set of parameter values these routines are run to determine the aircraft's aerodynamic characteristics, structural integrity and weight. By adjusting the parameters within a formal optimisation program, a minimum weight design that meets all performance requirements can be obtained. The mission requirements can also be varied to find optimum configurations over a range of vehicles from Micro UAVs (currently of interest to the US Department of Defence) through to larger, more conventionally sized vehicles for use in mineral exploration, search and rescue, surveillance and others.

4.4 Micro UAVs

An international competition to use Multi-Disciplinary Optimisation (MDO) techniques to design the smallest UAV capable of acquiring an image 600 m from the launching point, has initiated work on micro UAVs. This new class of UAVs, with wingspans potentially as small as 7 cm, presents challenges in airframe design (very low Reynolds numbers), propulsion and energy storage design, and sensor design. However, it creates a new range of applications in both military and civilian applications, amongst which include the biological and chemical sensing within confined urban environments.

5 UAV GROUP PERSONNEL

Dr K.C. Wong (Lecturer)

Research Interests: Design, development and operation of Remotely Piloted Vehicle (RPV) systems. Airframe, instrumentation and telemetry design. Low speed experimental aerodynamics.

Tel: 61-2-9351 2347

e-mail: kc@aero.usyd.edu.au

Dr Peter W. Gibbens (Lecturer)

Research Interests: Flight mechanics, nonlinear and robust flight control, state estimation and data fusion, parameter estimation.

Tel: 61-2-9351 7350

e-mail: pwg@aero.usyd.edu.au

Dan M. Newman (Lecturer)

Research Interests: Flight dynamics and mechanics, aircraft performance, engineering and developmental flight testing, data acquisition and analysis for aeronautics, aeroelasticity and dynamics of aeronautical structures.

Tel: 61-2-351 5527

e-mail: dan@aero.usyd.edu.au

Dr Doug J. Auld (Senior Lecturer)

Research Interests: Wind tunnel testing, low speed aerodynamics, flow separation effects, supersonic and unsteady flows, computational fluid dynamics. Aerofoil design and analysis. Development of Finite-element and Boundary-element fluid flow analysis methods. Flow prediction using Direct Molecular Simulation (DSMC) methods. Real-time control software and computer hardware development.

Tel: 61-2-351 2336.

e-mail: douga@aero.usyd.edu.au

Stuart Wishart (Postgraduate)

Research Interests: Application of parallel processing algorithms to the direct molecular simulation of gas flows (DSMC). Real-time GPS software development.

Tel: 61-2-351 7135

e-mail: swish@aero.usyd.edu.au

Hugh Stone (Postgraduate)

Research Interests: Aircraft configuration design and control optimisation.

Tel: 61-2-9351 2393

e-mail: hstone@aero.usyd.edu.au 

Jeremy A. Randle (Postgraduate)

Research Interests: Development and operation of UAVs.

Tel: 61-2-9351 7128

e-mail: jag@aero.usyd.edu.au

KuoShen Choong (Postgraduate)

Research Interests: Multi-rate sensor data fusion and GPS attitude determination

Tel: 61-2-9351 2393

e-mail: kuoshen@aero.usyd.edu.au

David P. Boyle (Postgraduate)

Research Interests: Nonlinear and robust flight control.

Tel: 61-2-9351 3040

e-mail: dpb@aero.usyd.edu.au

Philip W. Blythe (Postgraduate)

Research Interests: Neural networks, System Identification, Evolutionary Robotics, and Adaptive Control.

Tel: 61-2-9351 7138

e-mail: philby@aero.usyd.edu.au

6 CONCLUSION

UAV technology development has been an active area of research in the Department of Aeronautical Engineering at Sydney University since 1988. These research activities have advanced capabilities in all of the core UAV technologies of design, construction, system development, control and guidance to implementable levels. All of these capabilities and technologies have been demonstrated either in-flight or by simulation. Current activities revolve around coalescing these technologies into autonomous airborne vehicle systems.

The achievements to date place this department in a position to launch into mission specific UAV developments. Accordingly, the department is seeking involvement in, and industry partners for, mission specific UAV system developments. Specific UAV roles have been identified and their requirements analysed. The capabilities of this department cover the requirements of systems for nearly all identifiable UAV roles.

7 PUBLICATIONS

Some publications related to UAV research are listed to illustrate activities by people in this department:

[1] Boyle, D.P., Chamitoff, G.E. (1996). "High Performance Maneuver Control for Autonomous Flight Vehicles". Proc. AIAA Conference on Guidance, Navigation and Control, 1996.

[2] Stone, H., Wong, K.C. (1996). "Preliminary Design of a Tandem-Wing Tail-Sitter UAV Using Multi-Disciplinary Design Optimisation". 23rd Annual AUVSI Symposium (AUVSI '96) - 'Innovations for the Future' - 15-19 July 1996 - being published.

[3] Choong, K.S., Gibbens, P.W., Newman, D.M. and Wong, K.C. (1996). "Multirate Sensor Data Fusion and GPS Integration". Interim Report for DSTO Contract 332034 - April 1996.

[4] Chamitoff, G.E., Wong, K.C., Newman, D.M. and Boyle, D.P. (1996). "Application of Fuzzy Control for Airborne Vehicles". Final Report for DSTO Contract 340548, April 1996.

[5] Blythe, P. W. and G. Chamitoff (1995). "Estimation of Aircraft Aerodynamic Coefficients using Recurrent Neural Networks." In Proc. Second Pacific International Conference on Aerospace Science and Technology.

[6] Newman, D.M., Wong, K.C. (1995). "An Atmospheric Model for Small Remotely-Piloted Vehicle Simulation and Analysis". Second Pacific International Conference on Aerospace Science and Technology - Sixth Australian Aeronautical Conference 1995. Melbourne, Australia, 20-23 March 1995. The Institution of Engineers, Australia Conference Proceedings pp 579-584.

[7] Chamitoff, G.E., Wong, K.C. and Newman, D.M., Boyle, D.P. (1994). "Autonomous Guidance and Robust Control for the Recovery and Landing of a Remotely Piloted Vehicle". Technical Report Contract 340012, DSTO Report, December 1994.

[8] Chamitoff, G.E. (1994). "Autonomous Guidance for the Recovery and Landing of a Remotely Piloted Vehicle." Proceedings of the IFAC Aerospace Control Conference, pages 334-339, Palo Alto, USA, September 1994.

[9] Chamitoff, G.E., Ho, L.C. (1995). "Real-Time Autonomous Guidance for Self-Piloted Vehicles". Second Pacific International Conference on Aerospace Science and Technology - Sixth Australian Aeronautical Conference 1995, Melbourne, Australia, 20-23 March 1995. The Institution of Engineers, Australia Conference Proceedings pp 957-962.

[10] Newman, D.M., Wong, K.C. (1993). "Six Degree of Freedom Flight Dynamic and Performance Simulation of a Remotely Piloted Vehicle". Aero. Tech. Note 9301, The Department of Aeronautical Engineering, University of Sydney.

[11] Wong, K.C. (1993). "The Development of a Low-Cost Research R.P.V. System". PhD Thesis, The Department of Aeronautical Engineering, University of Sydney.

[12] Wong, K.C. (1993). "A Low-Budget Approach to the Development of a Research RPV System". 10th International RPV Conference, 28-31 March 1993, Bristol, U.K., Proceedings pp. 19.1-19.14.

[13] Gibbens, P.W. (1992) "Robust Nonlinear Control for Flight Manoeuvres" PhD Thesis, Department of Electrical Engineering, The University of Newcastle, NSW, Australia.

[14] Wong, K.C., Newman, D.M. (1989). "Exploratory Study into the Use of a Remotely Piloted Vehicle (RPV) for Aerodynamic Research", The Australian Aeronautical Conference 1989, Melbourne, 9-11 October 1989. Preprint of Papers pp. 27-30.

[15] Gibbens, P.W. and Goodwin, G.C. "Robustness Issues in Manoeuvre Autopilot Design", The Australian Aeronautical Conference 1989, Melbourne, 9-11 October 1989. Preprint of Papers pp. 38-42.


For further information please contact:

Dr K.C. Wong 

Tel: 61-2-9351 2347 Fax: 61-2-9351 4841

e-mail: kc@aero.usyd.edu.au

or any one in the UAV group.