Charles E. Hall, Jr.*

North Carolina State University

PO Box 7910

Raleigh, NC 27695

and

Siddhartha Mukherjee·

Glistening Web Communications Corporation

PO Box 994

Blue Bell, PA 19422

The World Wide Web is a tool that offers many new and exciting possibilities. The use of the World Wide Web for educational purposes has been implemented. These applications have opened new worlds for students, one area that has not been investigated is the use of the World Wide Web as a means for students to teleoperate experiments. This paper describes work that has been performed to allow distant students to operate experiments that are airborne in a remotely piloted vehicle over the web.

The use of the World Wide Web (WWW) as a teaching tool is reaching into more classrooms everyday. Most of the WWW sites utilized by students are essentially interactive computer programs or databases. There are a few sites on the WWW that permit the visitor to operate a piece of equipment and perceive the results. Many teaching and research laboratory facilities incorporate computers that reside on the Internet. These facilities can be incorporated into a web site that will allow undergraduate or secondary school students access to perform experiments.

At North Carolina State University an active program in Remotely Piloted Vehicles (RPV) has been in operation for the past twenty years. Initially this program was based solely on the capstone senior design course, but this has expanded during the past seven years to include a research based component. There has been a symbiotic relationship between the senior design and the research programs. Equipment and techniques developed for one program have been applied to the other. Examples of this symbiosis are the on board computers systems for flight control and data acquisition and telemetry systems.

This paper describes efforts to connect a web site to the computers systems that are operating on board a RPV during flight tests via a telemetry data link. Students are able to activate, operate and obtain data from experiments on board the RPV in nearly real time. The flight tests facilities, equipment and aircraft will be described. The WWW techniques will be presented. Finally, the integration of all the systems will be described.

The NCSU Flight Test Facility is located north of Butner, NC on university property which is surrounded by the North Carolina Department of Forestry land, and other unpopulated state and federal lands. This provides for an isolated and unpopulated area for the flight testing of RPV's. The Butner Flight Test Facility is located 28 air-miles north of the main NCSU campus. Due to the remoteness of the facility there are no Internet connections located on site. The flying site is located within the NCSU Beef Cattle Research Station, which does have telephone lines that are a mile from the runway. The runway is a 450 ft by 50 ft blacktop runway, oriented in the 03-21 directions. Located at the flying site is a computer operated weather station that measures temperature, pressure and humidity, and records this data for use in post flight analysis.

The Converse RPV (Fig. 1) was designed at NCSU for flight testing of small payloads. It was designed to minimize the effort and crew required for flight test operations. The Converse RPV is capable of carrying a payload of over 8 pounds in a space of 0.4 cubic feet. This is an ample space to contain the telemetry system, one or more of the Flight Computer Systems (FCS), various transducers and battery packs. There are RPV's at NCSU that are capable of carrying larger payloads, but for this demonstration and most teleoperated experiments the payload capability of the Converse RPV is adequate.

The computer on board the RPV is a key element for the system. The Flight Computer System was originally developed at NCSU in 1991 to be a general purpose on board computer system that is capable of data acquisition and implementation of control laws[1-6]. The system is based on the Dallas Semiconductor DS2250 microcontroller. The DS2250 is an 8-bit microprocessor, with 64K bytes of nonvolitile memory, two programmable timers and a serial port. A major advantage of using the DS2250 is the ability of programming it in a high level language using a Franklin, Ltd. ANSI C compiler. This permits the generation of complex code with ease, speed, and accuracy. The FCS can be programmed with code that is used multiple times, or it can be reprogrammed at the flying site. For example, in one afternoon 11 different student designed flight control laws were flight tested and performance data of their systems was recorded for each student. Included in the FCS are several support circuits. A radio interface receives signals from the radio control system and supplies signals to the servos which typically operate the control surfaces. There are six channels of analog to digital converters which typically operate with 13.5 bits of precision, but this is modifiable via programming. Also, due to the conversion technique two channels also generate a hybrid analog-digital integration of the signal. The analog to digital converters are calibrated against four precision voltages during operation to ensure accuracy of the converted data. For communications with other computers and devices there is a RS 232 interface.

To date NCSU has flown numerous transducers on board RPV's such as angular rate, pressure, temperature, accelerometers, velocity, position, strain gauges and angle of attack transducers to name a few. An FCS has been interfaced to a PSI module for measurement of multiple pressures. In the spring 1998, the 1997-1998 Senior Design program will fly a PSI module on a RPV with built in pressure taps in the wing and blended wing-fuselage portions of the aircraft. With this selection of transducers a wide variety of experiments can be designed for the students. In addition, the on board computers can control any device that is carried in the RPV, to include the RPV itself.

One of the devices that the RS232 serial port of the FCS has been connected to is a telemetry system. The telemetry systems at NCSU are b ased on commercially available radiomodems. The telemetry system primarily for the research efforts is a high speed, 128KB, full duplex, 900 MHz spread spectrum (SS) system. While the telemetry system for the senior design program is a 19.2KB, simplex, 450MHz, narrow band FM (NBFM) system. Both system have their inputs and outputs as standard RS-232 and can be operated in the asynchronous mode. The SS system can also be run in the synchronous mode. The other significant differences between these two system are size, weight and cost. At first is may seem that the NBFM system is severely limited by its 19.2KB simplex structure, but since the data is transmitted in a integer binary format that is compressed and not as ASCII numbers in engineering format the small bandwidth can transmit a large quantity of data. This compressed integer binary system has been employed both with the SS system and on board data storage systems to increase their bandwidth. The NBFM system has a key advantage over the SS system, in that operates at a lower rf frequency and thus the NBFM system experiences less atmospheric attenuation. Since both systems operate with a nominal 1W rf output, the NBFM system has a longer range. The NBFM system has a range that reaches from the main NCSU campus to the Butner Flight Test Facility a distance of 28 miles. The SS system has a maximum range, with acceptable bit error rates, of 4 miles. Thus, the SS system requires the use of a telephone line with modems to span the distance, since there are currently no Internet connections at the flying site. Initially both the NBFM and SS systems will use a telephone line connection from main campus to the flying site. To accomplish this, the Internet connected computer connects via a modem over telephone lines to a ground computer at the flying site. This computer acts as a switch between the modem and the telemetry system. Thus, linking the Internet access computer with the FCS on board the aircraft.

In the final form, the web server will be comprised of an Intel CPU utilizing the Linux operating system. Apache Web Server software will be used and supplemented with custom CGI scripts written in C and compiled using the gcc compiler together with the gnuplot package for data analysis and presentation. This arrangement results in web server response times that are much smaller than the desired maximum latency of the communication channel to the aircraft (100ms).

A student using a web browser accesses the teleoperated experiment web site at the appointed time. Eventually coordination will be via the web site which requires that the host web site be located at the flying site, or have a reliable connection with low latency to the flying site. This could, if necessary, be accomplished with a dial up connection to the flying site, or a leased line. Since placing an Internet connection at the flying site is in the future, coordination at this time is accomplished by voice using a cellular phone to call the student. This coordination allows the student to know when the RPV is airborne and it is in a position for the safe operation of their experiment.

Depending on the nature of the experiment various safeties are included on the aircraft that are removed by the flight crew on the ground and the last one by the pilot in the air. When the student activates the experiment this is done by either the operation of menu buttons or by the transmission of commands and data to the aircraft. This allows for the modification of the test conditions during a flight if the data demonstrates that the experiment requires modifications. Such menu based interfaces are readily converted into web based forms to accept, validate and transmit messages to the aircraft during flight. Communication between the student's web browser and the web server will be encrypted and protected against out of band attacks (such as session hijacking).

The host computer through the telemetry link communicates to the Flight Computer System on board the RPV. The FCS measures the values produced by the transducers, activates items and provides inputs, as required, to the control surface actuators. Data obtained by the FCS is transmitted to the host computer and relayed to the student. The data is displayed to the student in a graphical format and supplied as a numerical data array for further analysis. Again these data and graphs will be translated into a web page for easy viewing with the student's browser.

The goal of this work was to demonstrate the feasibility of the use of the World Wide Web as a means for teleoperating experiments. The was no a specific experiment in place for this project. This section will describe two possible experiments that could be implemented utilizing this technique.

One application would involve the student taking measurements during flight to determine the elevator angle to trim curve of the aircraft. The transducers required for this would be a pitot-static tube, an angle of attack vane and elevator deflection. Data would be taken during portions of straight and level flight. The weight of the aircraft would be determined prior to flight. With these measurements and the atmospheric conditions provided by the ground based weather station, the student would be able to calculate the elevator angle to trim curve and trim lift curve.

Another application is the combined use of a wind tunnel and flight tests to demonstrate the lift generated by an airfoil. First the student would teleoperate surface presssure scans of the RPV airfoil in a low speed, subsonic wind tunnel. These surface pressure scans would be performed by the use of a PSI module. Teleoperating the experiments in the wind tunnel would be performed as it would be with the aircraft, except for the web server would be directly connected to the PSI module at the wind tunnel site. Thus simplifying the system. The pressure distributions for a range of angle of attacks would be integrated by the student to calculate the lift generated by the airfoil. The data would also allow the student to generate the lift curve of the airfoil. The RPV would have pressure taps built into the airfoil. The PSI module would be placed into the aircraft and the web connection would be made by the telemetry link. The student would command surface pressure scans at different airspeeds of the RPV. The student would be able to integrate these pressure distributions to obtain the lift at each speed condition. These lift and pressure distributions would then be connected together, thus demonstrating to the student the basic concept of flight.

These example experiments are intended to provide all the possibilities that are available. It is anticipated that prior to conducting the teleoperated experiments, the students, or their teacher, would have a video based lecture available to explain the goals and concepts of the experiments. Thus, any secondary school with Internet access and a TV/VCR can have access for their students to state-of-the-art experimental facilities.

This project has been an unfunded cooperative effort between the authors. Due to this situation, the project could not be completed due to a crash in August 1997 of the Converse RPV. During a flight test of the Converse RPV, while in a low energy state, flew into the downwash generated by the wind coming over a forested area near the flying site. The Converse was underpowered during take-off and go-around landing aborts, but acceptable for other flight phases. The Converse RPV was repaired and modified to include a larger engine. The repair and modifications returned the Converse to flight status in December 1997, but it has not completed its requalifying flights at the time of this writing.

The interfacing of Internet connected computers to the telemetry system, and the resulting telemetry link to the FCS has been completed. The ability to control the actions of a FCS over the net/telemetry link has been demonstrated and qualified for flight testing. Web servers and associated web pages have been created.

It is planned that initial flight tests will be conducted during the spring of this year. The experiments utilized for these initial flight tests will not be based on aerodynamics, but rather they will be directed at determination of a limit of the bit error rate of the system. This will have the FCS sending various predetermined data streams when directed by the teleoperated experimenter. The system will be programmed to compare the string received on the ground to the string that should have been sent. After the initial testing, a test group of students will operate an experiment on board the RPV.

1. Hall, C. E. Jr., " Microcomputer Application for Small-Scale Flight Vehicle Autopilots/Stability Augmentation Systems", Proceedings of the ISMM International Conference on Engineering and Industrial Applications of Microcomputers, pp. 111-114, Long Beach, CA, December 1991.
2. Hall, C. E. Jr., "An Airborne and Ground-Based Computer Network for Command, Control and Data Acquisition in a RPV Flight Test Program", ISCA International Conference on Computers in Engineering and Medicine, pp.231-234, Indianapolis, IN, March 1996.
3. Hall, C. E. Jr., "A Distributed-Redundant System for Autonomous UAV's", ISCA International Conference on Computer Applications in Industry and Engineering, pp.291-295, Honolulu, HI, December 1993.
4. Hall, C. E. Jr., "A Stability Augmentation System for Student Designed remotely Piloted Vehicles", AIAA Aircraft Design Systems Meeting, AIAA 92-4261, Hilton Head, SC, August 1992.
6. Fontenrose, P. L. and Hall, C. E. Jr., "Development and Flight Testing of a QFT Pitch Rate Stability Augmentation System", Journal of Guidance, Control and Dynamics, Sept-Oct 1996, pp.1109-1115.
7. "F/A-18E/F Simulation Support Critical Design Review", NCSU Flight Research Group, Report to the Naval Air Warfare Center, Patuxent River, MD, March 1995.

Figure 1. The Converse RPV.