Development and Testing of the X-2C Unmanned Aircraft System for the 2009 AUVSI Student UAS Competition

Transcription

1 Development and Testing of the X-2C Unmanned Aircraft System for the 2009 AUVSI Student UAS Competition Wade Spurlock, Team Lead Daniel Wilson, Avionics Lead Marty Brennan & Travis Cope, Fall & Spring Airframe Leads Abstract The 2009 Student UAS Competition hosted by AUVSI marks X-ipiter s sixth year of participation. The team has developed a modified Vee systems engineering model, resulting in a robust student-designed-and-built airframe carrying avionics of commercial off-the-shelf hardware and studentwritten software. X-ipiter s systems engineering approach addresses the requirements and design of the separate subsystems, coding of software and fabrication of the hardware with subsystem testing, and integration of components into the final product. X-2C is divided into four main subsystems: Guidance, Navigation, and Control (GNC); Imagery; Airframe; and Safety. The GNC subsystem consists of the Piccolo LT autopilot coupled with a laser altimeter and magnetometer allowing autonomous takeoff, waypoint navigation, and landing. The onboard components of the imagery subsystem consist of a Sony D70 analog pan/tilt/zoom camera, a video IP server to digitize the video, and an Ethernet bridge to link with the groundstation. On the ground, student-written software is used to gather and analyze imagery to determine the target parameters. The avionics are housed in a student-designed-and-built airframe fabricated from preimpregnated carbon/fiberglass hybrid composites. The airframe is capable of carrying a payload of up to 25 lbs, and the X-2C UAS gross takeoff weight is 55 lbs. The safety subsystem key element is the Dual Power Servo Interface Twin Maxi, which allows servo power redundancy and inputs from both the autopilot and the manual backup R/C receiver. While X-ipiter is returning with generally the same X-2C airframe from the 2008 competition, new onboard avionics improve performance for the Intelligence, Surveillance, and Reconnaissance mission. 1

2 Introduction X-ipiter s X-2C UAS combines years of experience with a systems engineering approach for the 2009 Student UAS competition. The team used many systems engineering influences and references to develop a modified Vee model, seen below in Figure 1, with five phases to progress from the concept of operations and defining the mission requirements to full system integration and flight demonstration. Figure 1. X-ipiter's Systems Engineering "Vee" Model Phase I: Concept of Operations The concept of operations is for X-ipiter s X-2C Unmanned Aircraft System (UAS) to fly an Intelligence, Surveillance, and Reconnaissance (ISR) mission to perform aerial support for a simulated Marines combat operation. Phase II: Requirements Analysis, Subsystems Functional Analyses, and Component Detail Design Phase II of X-ipiter s design process for X-2C encompasses the development of the requirements, subsystems, and components. After Phase II was completed, a two-part presentation in the form of a Conceptual/Preliminary Design Review was given with an invitation to the entire Bagley College of Engineering at Mississippi State University. The full system requirements are presented in the statement of work (SOW) for the Student UAS Competition. The major requirements and constraints that were determined to impact the design and performance of the UAS are listed below. Requirements The system shall be capable of autonomous flight. The system shall be capable of real-time imagery. The system shall be capable of target identification. The system shall be capable of safe operation. 2

3 Supplemental requirements The system shall be capable of autonomous take-off and autonomous landing. Constraints The system shall be capable of avoidance of the competition specified no-fly boundaries. The system shall be capable of remaining in flight between MSL. The mission shall be completed in a maximum of 40 minutes. The system shall have a maximum gross takeoff weight of 55 lb. The system shall have a maximum airspeed of 100 knots. The system shall be capable of operating within specified environmental conditions. Subsystem Functional Analyses: A functional flow block diagram (FFBD) was constructed from the concept of operations requirements analysis. The FFBD describes a functional flow of the X-2C UAS operations from three levels: system level, aircraft level, and mission level. The system level diagram in Figure 2 below describes the broad steps from requirements analysis to the concept demonstration. The aircraft level diagram in Figure 3 details the flight demonstration from completing pre-flight checklists to landing and post-mission briefing. The mission level diagram in Figure 4 lists the specific tasks to be performed during the autonomous portion of the mission. 3.0 Fabricate Airframe 1.0 Define System Requirements 2. 0 Design Unmanned Aircraft System AND 4.0 Procure COTS for Avionics Subsystems 5.0 Perform Systems Integration and Test & Evaluation 6.0 Perform Concept Demonstration Phase (Student UAS Competition ) Figure 2. System Level Functional Flow Block Diagram 6.0 (Ref) Perform Concept Demonstration Phase (Student UAS Competition) 6.1 Perform Pre-flight Checklists and Procedure 6.2 Perform Autonomous Takeoff and Climbout 6.3 Perform Autonomous Flight Mission 6.4 Perform Autonomous Approach and Landing 6.5 Perform Post-Flight and Post-Mission Operations Figure 3. Aircraft Level Functional Flow Block Diagram 3

4 6.3 (Ref) Perform Autonomous Flight Mission Fly to Secured Corridor Ingress Waypoint AND Navigate Waypoint En Route to Search Area Search for and Identify Target(s) along Secured Corridor OR Fly Programmed Flight Path Fly to Emergent Waypoint Search for Targets Search for Pop-Up Target(s) Downlink Real- Time Video to Ground Station Downlink Real- Time Video to Ground Station Identify Targets and Target Characteristics Identify Pop-Up Target(s) and Target Characteristics Figure 4. Mission Level Functional Flow Block Diagram 6.4 (Ref) Perform Autonomous Approach and Landing Based on the mission requirements, the function of the system is split into the main subsystems of Guidance, Navigation, and Control (GNC); Imagery; and Airframe. Additionally, the safety of the X-2C system is an important factor and is included throughout the design, fabrication, and integration. Component Detail Design Guidance, Navigation, and Control A trade study was used to evaluate several commercial off-the-shelf (COTS) autopilots, including the UAV Navigation AP04R, Procerus Technologies Kestrel, Micropilot 2028g, Paparazzi Tiny v2.11, and Cloud Cap Technology Piccolo LT autopilots. Determining factors were based on the requirements and GNC subsystem functional analysis. The GPS, return home, flight termination, and weight factors received weights between 2 and 3 as these are desirable features but not as important as autonomous takeoff/ landing and cost which received weight factors of 4 and 5, respectively. The cost category is an opportunity cost taking into account both monetary and time expenses. Scores for each choice were assigned on a scale from 1-5 with 1 being the worst score and 5 being the best. Table 1 details the results of the autopilot trade study. After consideration of each of the factors in the trade study, the Piccolo LT autopilot system by Cloud Cap Technologies was chosen for the GNC subsystem of the X-2C UAS. Table 1. Guidance, Navigation, and Control Subsystem Trade Study Choice UAV Navigation Procerus Micropilot Paparazzi AP04R Kestrel 2028g Tiny v2.11 Factor (Weight) Cloud Cap Piccolo LT GPS Capable (3) 3 x 3 3 x 3 3 x 3 2 x 3 4 x 3 Auto Takeoff/Landing (4) 3 x 4 2 x 4 3 x 4 1 x 4 4 x 4 Return Home (2) 4 x 2 4 x 2 4 x 2 1 x 2 5 x 2 Flight Termination (3) 4 x 3 4 x 3 4 x 3 1 x 3 4 x 3 Weight (2) 3 x 2 4 x 2 5 x 2 2 x 2 4 x 2 Cost (5) 1 x 5 4 x 5 3 x 5 2 x 5 3 x 5 Total

5 Imagery The system s real-time imagery requires a camera and communications link to the groundstation, where an operator is controlling the camera and viewing the video. During the mission, targets lying off the flight path and aircraft attitude reduce the effectiveness of a fixed camera. A pan/tilt/zoom camera was chosen for better control of the field of view while gathering imagery. There were several options for the team to choose from concerning the camera to be used in X- 2C. To make this decision, a trade study was conducted to evaluate each camera based on certain predetermined criteria. Table 2 below lists the different cameras considered, the criteria by which they were evaluated, and the ratings each camera received for each criterion. As a result of this trade study, the team added the Sony EVI-D70, an analog pan/tilt/zoom camera with 18x optical zoom and an effective resolution of 768 x 494. Table 2. Imagery Subsystem Trade Study Choice Sony D70 Toshiba IK-WB21A Bronzepoint PTZ 270 Factor (Weight) Cost (2) 5 x x 2 5 x 2 Performance (2) 4.2 x x x 2 Resolution (3) 3.5 x 3 5 x x 3 Zoom (2) 4.2 x x 2 5 x 2 Size (1) 3.9 x 1 5 x x 1 Gimbal System (1) 4.2 x x x 1 Total In previous years, the transmission of analog signals was susceptible to a significant amount of noise, and digital cameras with built-in conversion experienced severe lag issues. To combine the advantages of both analog and digital cameras, an analog camera in conjunction with an external video conversion device was specified for the subsystem. X-2C contains many devices that must communicate with operators on the ground. The number of transceivers is minimized to reduce the amount of interference. The two features in the imagery subsystem that require communication are the video feed and camera control. The external video conversion device must then interface with a multi-device bus capable of supporting a security feature and handling the video and camera control traffic. Before software is designed for the groundstation aspect of the imagery subsystem, it is appropriate to determine basic roles for the operators. These roles are assigned according to the amount of attention that certain tasks demand. As a result, one camera operator controls the pan, tilt, and zoom of the camera and features such as recording the video and taking still-shots to be analyzed later. Another operator analyzes the images and obtains required target information such as location, size, and shape. To maximize efficiency of this subsystem, specialized software will be designed to make each operator s job as efficient as possible. The camera control software must send commands to the camera to control pan, tilt, and zoom. It must also be able to capture and view video and save individual snapshots along with the necessary information to analyze the imagery for target parameters. The image analysis software must load images and the corresponding data for plane and camera states. The user must be able to determine an object s GPS coordinates, size, and orientation. 5

6 Airframe Team X-ipiter has taken an evolutionary approach to developing a reliable UAS for use in competition as well as for real-world applications. The design of X-2C has evolved from Senior Telemaster, X-1, X-2A, and X-2B. The airframe used for the 2004 AUVSI UAV Competition was an off-the-shelf model, which made construction easy. It featured a tail dragger configuration and a split horizontal stabilizer. The engine was in a tractor configuration and required glow fuel. Despite its simple construction, the Senior Telemaster had insufficient internal room for the payload. To provide the payload volume required by the avionics system components, Team X-ipiter designed X-1 for the 2005 UAS competition. This airframe was fabricated using fiberglass with a wet lay-up fabrication technique and featured a tricycle landing gear configuration and gasoline engine in a tractor configuration. The internal volume greatly increased from that of the Senior Telemaster; however, new challenges with the engine placement included restricted access to the payload and exhaust residue interference with the camera surveillance system. The X-2A airframe was used in the 2006 competition to address problems associated with the X- 1 air vehicle. X-2A was constructed of preimpregnated carbon fiber using student-built carbon fiber molds, simplifying the manufacturing process and helping to decrease the gross weight of the air vehicle. The engine was placed in a pusher configuration, eliminating surveillance interference and allowing greater access to the payload areas. While problems associated with the X-1 air vehicle were solved, a high cruise speed inhibited the surveillance capability of X-2A. Built for the 2007 competition, the X-2B airframe evolved from the design of X-2A. To decrease the minimum flight speed and improve stability, the wingspan and boom lengths were increased. The fuselage length was also increased resulting in a greater internal volume. A newly installed brake system minimized the aircraft s landing distance and ensured safer ground taxi operation. The X-2C air vehicle has an improved design from X-2A and X-2B. Although the basic twinboom pusher design was retained, X-2C s performance was enhanced in the areas of maneuverability, structures, and vibration isolation of avionics components. Improvements were also made to the landing gear and nose cone. The desired cruise speed of 45 knots and slow flight characteristics of X-2B led to the wing design remaining mostly unchanged for the X-2C air vehicle. In order to enhance slow-flight maneuverability, the chord of the control surfaces on the wings was increased. The SD7062 was selected once again for the wing airfoil. This airfoil is a high lift airfoil in low Reynolds numbers and has shown great reliability in past Xawk systems. After examining the payload components which would be added to the system, the fuselage size was chosen. The wing and fuselage sizes can be seen in Table 3. A square cross section was selected due to ease of fabrication and effective use of internal space. Wings Fuselage Table 3. X-2C Wing and Fuselage Dimensions Parameters Value Span 128 in Chord 16 in Area ft 2 Aspect Ratio 8 Control Surfaces 4 in x 21 in Length 45 in Width 9 in Height 9 in Internal Volume 2.11 ft 3 The vertical and horizontal tails were initially designed based upon the wing sizing using the tailvolume coefficient method. Tail volume coefficients were determined by historical guidelines for general 6

7 aviation aircraft and past X-ipiter designs. A static and dynamic stability analysis was performed to calculate the precise sizing and placement of the empennage. With the tail placed 55 inches behind the wing, the final sizing and tail volume coefficients are presented below in Table 4. Table 4. X-2C Empennage Dimensions Parameter Verticals Horizontal Height/Span 12 in (each) 33 in Chord 9 in 9 in Area 216 in 2 (total) 297 in 2 Aspect Ratio Tail Volume Coefficient Safety An integral part of the system requirements is safety. As specified by the requirements, the weight of the airframe shall have a maximum gross takeoff weight of 55 lb. This gross weight will include all system components and fuel. A manual override system is required to address a communication or hardware problem associated with the GNC subsystem. In this event, a safety pilot will control the system. If at any point in flight the system goes beyond control of the autopilot, it is extremely important that the system respond in some way to correct. The corrective action will consist of a primary operation to regain control and a secondary operation if control of the system cannot be regained. Three main factors will be considered when selecting the propulsion system for the X-2C UAS: power requirement, power-to-weight ratio, and reliability. Reliability was the foremost concern. An unintended shutdown of the engine during flight could lead to a rough landing, potentially damaging both the aircraft and the equipment on board. Given the maximum gross weight requirement, a propulsion system with a large power-to-weight ratio is desirable. A minimum of 6 hp is required for X-2C. Safe operation of the system includes the use of a safety pilot, ignition of the propulsion system, airspeed, and battery colors. The X-2C UAS will have the capability of being flown by a safety pilot through the autopilot or completely bypassing the autopilot by using the backup R/C receiver. Starting the propulsion system of X-2C could potentially be dangerous in two regards: the act of turning the prop is dangerous to the starter, and the system could accidentally be started. To mitigate the risk associated with igniting the propulsion system, a supplementary ignition system will be present and a series of switches will be required to be activated before the engine will crank. Airspeed affects not only the flight characteristics of the system but also the ability of the imagery system to collect clear pictures of the targets. Maximum airspeed is also provided within the requirements analysis and constraints. The system will be designed for a relatively slow airspeed to utilize the imagery and to remain stable. The radio system should be designed to allow the aircraft to fly as far as possible. Using the operation area coordinates from the 2008 AUVSI competition as our guide, the farthest distance from the groundstation that our aircraft could travel was determined to be approximately 2,785 feet. Including a safety factor of 2, the radio system should be designed for a range of approximately 1 mile. Interference may disrupt the use of devices in two ways: noise may be interpreted as commands or commands may be distorted and interpreted incorrectly. This can cause unexpected, and possibly dangerous, behavior of the system. To prevent interference to surrounding devices within the airplane, it is important to use twisted and shielded wires. The size of the wiring is also an important consideration to safely operate under operational loads. In addition to interference, faulty connections can also cause major problems. If a critical component loses power or signal, this can also cause problems similar to those described above. To help prevent faulty connections, proper connectors and switches are used. 7

8 Due to forces on the system, all components must be fastened securely. Insufficiently securing components may result in unexpected behavior of the system. As a result, component fasteners will be selected to ensure safe and reliable assembly. Phase III: Code Software; Fabricate Hardware; Component Testing & Evaluation; Subsystem Integration, Testing, & Evaluation Once the detail design of the subsystems and components is completed, Phase III begins with component fabrication and software coding. The components are tested and developed into the subsystems, which are also tested prior to full system integration. The full system is then evaluated with a series of ground and flight tests. In April, X-ipiter participated in the AIAA Region II Student Conference in Huntsville, AL, and presented the Critical Design Review. Guidance Navigation and Control Autopilot The Piccolo LT autopilot system by Cloud Cap Technology is the foundation for autonomous flight in the X-2C UAS. The Piccolo LT, seen below in Figure 5, communicates with the operator through the Piccolo Command Center, which provides information on aircraft location and orientation, indepth telemetry, and systems data. The operator uses this interface to supply the autopilot with waypoint information when programming the flight path. The autopilot also has several built-in safety features, including an R/C failsafe mode and a return-home command. Figure 5. Piccolo LT Autopilot Magnetometer and Laser Altimeter In order to meet the requirement of autonomous takeoff and landing, additional sensors are needed to supplement the autopilot system. As a standalone unit, the Piccolo LT uses GPS readings for heading reference and barometric data for altitude. While generally giving sufficient data for flight, these heading and altitude readings are not accurate enough for autonomous takeoff and landing. In order to enhance the Piccolo LT telemetry, a magnetometer and laser altimeter sensors were added. The HMR2300 by Honeywell is a three-axis digital magnetometer, seen below in Figure 6, that supplies a very accurate heading reading to the autopilot system. This allows the aircraft to follow a specified heading down the runway for takeoffs and landings. The laser altimeter by Latitude Engineering provides extremely accurate above-ground-level altitude readings during flight shown in Figure 7. The altimeter has a range from 10 cm to 1 km with a resolution of 10cm. Together, the precise measurements from these sensors give the autopilot system the capability of autonomous takeoff and landing. 8

9 Figure 6. Honeywell HMR2300 Magnetometer Figure 7. Laser Altimeter by Latitude Engineering DPSI Team X-ipiter chose to employ the Dual Power Servo Interface (DPSI) Twin Maxi V2.0 Pro, seen below in Figure 9, manufactured by Emcotec Corporation of Germany for the following reasons: Increased safety during flight testing Capability of operating multiple servos to a single channel Capability of increasing flight time Smaller Size and Weight than previous DPSI Twin (Figure 8) Figure 8. DPSI Twin: 7 x 3.5 x 0.6 in, 0.5 lb Figure 9. DPSI Twin Maxi: 3.5 x 2.4 x 1.6 in, 0.35 lb The DPSI system increases flight safety by allowing a secondary (backup) receiver to control the aircraft in the event of an emergency. The DPSI allows only the primary receiver to send commands to the servos. The secondary receiver may be operating at the same time as the primary, but the secondary receiver s commands will be ignored as long as the primary receiver has not tripped a fail mode. The DPSI servo outputs have a higher current limit than most receivers, allowing multiple servos to be connected to one output channel. The Xawk design requires two aileron, two flap, and two rudder servos. Therefore, three channels must have at least two servos connected to it. In addition, the nose gear is connected to the rudder channel for standardization purposes and reduction of the number of independent channels necessary. The DPSI Twin Maxi V2.0 Pro comes with a built-in power station that allows up to two batteries to be connected at a time. The batteries are internally paralleled. This configuration allows for battery power redundancy and greater battery lifetime. The batteries chosen to power the servos of the Xawk are 2-S Lithium Polymer battery packs (7.4 V) with a capacity of 3300 mah. This gives the aircraft 6600 mah of total power dedicated to the servos, allowing for a total of 1.5 hours of continuous use, well above the total flight time of the mission. 9

10 Backup Receiver A backup R/C system is employed so that control of the aircraft can be regained in the event of a failure of the autopilot. This system was required to be integrated with the current Futaba T9CAP Super radio transmitter because the Piccolo groundstation can interface only with Futaba radio systems. Using a radio system that cannot be integrated into the current Futaba radio system would require the safety pilot to operate two different radio systems. The R/C system chosen for this task was the Futaba R608FS and the TM-8 transmitter module, both from the Futaba Advanced Spread Spectrum Technology (FASST) product line. The FASST products are a line of 2.4 GHz spread spectrum radio technology designed for higher reliability. Three advantages offered by the FASST radio modules and receivers are the different radio frequency than the autopilot system, dual-diversity antennas, and frequency hopping. All of these features make the backup radio system more robust than in past Xawk systems. Guidance, Navigation, and Control Subsystem Integration In order to integrate the magnetometer and laser altimeter sensors with the autopilot, each sensor had to be installed and calibrated. These sensors take advantage of the Piccolo LT s two expandable payload ports. Once calibrated, the magnetometer was tested by turning the aircraft in a specified direction and comparing the magnetometer readings to a compass. The results for this test, shown in Table 5, demonstrate the magnetometer giving extremely accurate heading measurements crucial for autonomous takeoff and landing. Table 5. Magnetometer Testing Pass # Compass Heading ( ) Magnetometer Heading ( ) Error ( ) Similarly, the laser altimeter had to be calibrated to reflect the bias of being mounted a specified height off the ground. For the X-2C aircraft, the laser altimeter was mounted approximately 1 ft off the ground. Once calibrated, the laser altimeter was tested by leaving the aircraft level on the ground and pointing the laser altimeter at a wall. The aircraft was moved progressively farther from the wall to encompass a range similar to that seen in autonomous landing. The values returned by the laser altimeter were compared to the actual distance of the wall. The results of this test, shown in Table 6, demonstrate the laser altimeter giving very accurate altitude measurements crucial for autonomous takeoff and landing. Table 6. Laser Altimeter Testing Pass # Actual Height (ft) Altimeter Reading (ft) % Error Guidance, Navigation, and Control Testing The first step in the testing and evaluation of the autopilot system is performing a software-in-theloop (SIL) simulation. Cloud Cap provides the Athena Vortex Lattice (AVL) program that uses a vortexlattice method to determine stability derivatives in order to estimate the flight characteristics of an aircraft model as seen in Figure 10. Information from AVL is loaded into the Cloud Cap Simulator where a 6-10

11 DOF simulation of the aircraft is performed. This test is crucial to becoming familiar with operating the autopilot, validating the aircraft model, and tuning control loop gains. Figure 10. AVL Model Once several SIL simulations are completed, the next step is a hardware-in-the-loop (HIL) simulation. The simulator interfaces with the autopilot to provide simulated sensor data while the aircraft flies a simulated mission. The aircraft s control surfaces and telemetry perform as they would in flight but with the safety of testing the mission on the ground first. The last step in testing and evaluation of the autopilot system is flight testing. This is done by allowing the R/C pilot to take the aircraft into a flight pattern, engage the autopilot, and verify the operation and handling characteristics of the autopilot. Control loops are tuned until the aircraft performs as desired. Once tuning is complete, mock missions are flown and each segment from autonomous takeoff to autonomous landing are verified. Post-flight performance analysis is done using a MATLAB program supplied by Cloud Cap that displays graphically sensor data from the entire flight. This allows the user to compare control loop commanded inputs (green) to the aircraft response (blue) as seen in Figure 11. This aids in more effective tuning of the autopilot gains. Figure 11. Autopilot Response to Commanded Input 11

12 Radio Range Testing All radio systems were tested by simply moving the radio transceivers apart until the radio transmission was degraded to an unacceptable level. The aircraft was on the taxiway at Starkville-Bryan Airfield, and the radio transmitters were down the taxiway. In this configuration, a significant amount of the radiated signals is dissipated by the ground; therefore, these tests will measure an attenuated signal. The results of these range checks will be closer to a worst-case scenario. In the case of the backup radio transmitter, an attenuator was also used, as it was available for this transmitter. Unacceptable levels were different for the different radio systems. For the autopilot system, the primary consideration for an unacceptable level is loss of manual control. The groundstation was moved away from the aircraft while smoothly moving the surfaces of the aircraft. Jittery response was observed at approximately half a mile from the aircraft. In addition, the receive signal strength indication (RSSI) and acknowledgement ratio were observed at various distances from the aircraft. The plot of these values versus distance from the aircraft is shown in Figure 12. Figure 12. RSSI and Acknowledgement Ratio When testing the camera transmitter, the primary consideration for unacceptable level is loss of video signal. The RSSI value is a second signal strength parameter to consider. The camera transmitter will be tested in the same manner as the autopilot transmitter. When testing the backup radio transmitter, the radio system manufacturer s recommendations were used to determine acceptable range. The manufacturer recommends that the radio system be able to smoothly control the aircraft at a minimum of 112 feet while employing the attenuator. The system was maintained smooth control of the aircraft at 150 feet, which is well above the manufacturer s recommendations. Imagery Subsystem Ethernet was chosen as the communication protocol for the imagery subsystem. This standard can support high-speed communication that is sufficient to transmit quality video back to the ground. To bridge the Ethernet connections between the plane and the ground, the Microhard VIP2400 wireless Ethernet bridge was chosen. This bridge is capable of up to 54 Mbps and supports basic wireless network encryption. To work in conjunction with the Sony EVI D70 camera, seen in Figure 13 below, the Axis Communications RHP-IPVS Video Server is used to convert the video signal and transmit it over 12

13 Ethernet. Not only does this server, seen in Figure 14 below, digitize the analog video stream, but it also allows for remote control of the camera. Figure 13. Sony EVI D70 Camera Figure 14. Axis Communications Video Server The two student-written applications for the imagery subsystem are X-ipiter Camera Control Software (XCCS) and X-ipiter Base Station (XBS). XCCS, as its name implies, is responsible for the camera control and video aspect. XBS is responsible for analyzing the snapshots for target parameters. X-ipiter Camera Control Software (XCCS) XCCS, below in Figure 15, is a student-written application used for video capture and camera control. XCCS is written in Java, a platform-independent language which makes it possible to transfer the application to any system. To use XCCS, the user inputs the network address of a camera from which to import the video. Once video capture has begun, individual snapshots can be saved to a user-specified folder. In addition to capturing video, XCCS also communicates with another server for the airplane telemetry such as location, altitude, heading, pitch, roll, and yaw. This information and the camera s pan, tilt, and zoom are saved in a text file along with each snapshot, allowing XBS to make its calculations later on in the target identification process. Another important feature is the camera control. Based on keys pressed on the keyboard or joystick input, the camera s position can be changed to fix the view on an area of interest. Figure 15. Screenshot of XCCS X-ipiter Base Station (XBS) The X-ipiter Base Station (XBS) software is a student-written application used for target identification, location, and characterization. The XBS user interface, seen below in Figure 16, is also 13

14 implemented in Java for inter-operability. A picture can be selected from a list containing all of the images available in a shared folder between the XBS laptop and the XCCS laptop. The mathematics behind XBS is based on photogrammetry, the science of obtaining information about the real world from photography and other measurements. XBS begins by establishing a series of six coordinate systems that fully describe the position and orientation of the aircraft and the camera: local ground-parallel system with axes parallel to the standard directional axes aircraft wind axis system that takes into account angle of attack and sideslip aircraft body system camera base system for pan camera body system for tilt final photograph system with the x-axis out the right side of the image and the y-axis out the top of the image for the photogrammetric calculations themselves. All systems are assumed to originate at the center of gravity of the aircraft. A series of five threedimensional coordinate transformation matrices are developed by calculating a series of rotation angles from one coordinate system to the next. These matrices are then multiplied together to give a single resultant transformation matrix from the ground-parallel system directly to the photograph system. Since all photogrammetric calculations are in terms of the angles of tilt, swing, and azimuth, a second set of calculations are done to convert the resulting angles to the final tilt-swing-azimuth system. Information about the target, such as GPS location, length, and heading can be derived using the plane-state data obtained from the autopilot at the time the picture was taken. Finally, this information can be saved for later analysis by assigning the target a unique number. Figure 16. Screenshot of XBS Imagery Testing During a subsystem test, both the onboard and groundstation components of the imagery subsystem operated satisfactorily. The time that it took the video to reach the screen on the computer was measured by waving an object in front of the camera and waiting for it to move on the computer screen. This time was less than 1 second. Also, the camera command lag was measured by sending a command to change the pan and tilt of the camera and measuring the time it took for the camera to actually move. This measurement was also less than 1 second. While flying, both of these lags were measured by 14

15 sending a movement command and waiting for the change to be seen on the screen. This time was also less than 1 second. The field of view of the camera is an important characteristic, especially in the analysis of individual photographs so that size and direction can be calculated accurately. For this reason, the field of view of the camera was measured for different zoom settings by placing the camera perpendicular to a wall a known distance away. A tape measure was placed against the wall to measure the physical width of the image at each zoom setting. Basic trigonometry is used to calculate the angle spanning the perpendicular distance. Taking this angle and dividing it by the number of pixels across the measurement gives the average angle per pixel ratio. Multiplying this by the total number of pixels in a certain direction gives the field of view for that direction. Performing this calculation at many different zoom levels allows for a curve fit. The calculated pixel resolutions were plotted against their corresponding zoom settings. Figure 17 below clearly shows the point at which the camera transitions from optical zoom to digital zoom. Additionally, each region has been fitted with a different curve. These equations are displayed next to the trendlines for which they represent. Figure 17. Sony EVI D70 Field-of-View Measurements Airframe Fabrication The X-2C airframe draws many of its major performance, stability and payload capacity advantages from its rather modern fabrication process and materials. At the heart of each wing is a COTS 1.5in diameter carbon tube encasing a solid nylon rod, placed at the quarter chord station. In addition, a COTS 0.25in solid carbon rod is used as a rear spar to minimize torsional forces on each wing. This rod is positioned 9.5in from the leading edge of the wing. The wing's shape is ensured by several wing ribs, each made of 0.25in birch wood sandwiched between multiple layers of preimpregnated carbon/fiberglass hybrid referred to henceforth as hybrid. The two-part upper and lower wing skins are made of the same hybrid material with a solid Divinycell foam core. X-2C wing ribs were joined to the wing skins using a single piece called a pi-clip to distribute loads to and from the ribs more effectively than with the quarter-inch edge bond used in previous airframes. Right-angle carbon fiber stock was used within the fuselage to join the bulkheads to the fuselage skin. Longerons on the left and right sides of the fuselage were also added to increase the rigidity of the fuselage and provide a convenient location on which to mount payload components and run wires. 15

16 The fuselage fabrication was completed much of the same way, using the similar sandwich construction of hybrid and a Divinycell foam core. The three top hatches and nose cone are composed solely of hybrid. Three bulkheads separate the fuselage into a nose gear compartment, main payload compartment, an autopilot compartment, and a rear compartment. Each bulkhead is composed of a sandwich construction of 0.25in oak and hybrid material. Segments of carbon-fiber right-angle stock join the bulkheads to the flat sides of the fuselage. Longer segments of the right-angle stock are used as longerons on the port and starboard sides of the fuselage. Besides greatly increasing the rigidity of the fuselage, they provide a convenient location on which to securely mount payload components and neatly run wires. Two wing-fuselage attachment stations, affixed to the second and third bulkheads, secure the wings to the fuselage in a male/female carry-through method using locking pins. A new trailing link shock-absorbing nose gear system is used to damp minor shocks during taxi testing and to absorb large landing loads without any loss in ground maneuverability. A removable nose cone is added to access the area forward of the front bulkhead, allowing nose gear maintenance and battery placement the greatest distance forward. The removable nose cone is similar to nose cones of previous X-2 air vehicles but securely attaches to the fuselage with two screws and nut plates on each side. The empennage of the X-2C air vehicle is mounted to the wings by twin booms and consists of twin vertical stabilizers joined by a horizontal stabilizer in a pi configuration. Designed specifically for large model aircraft, Kroma engines deliver the high performance required for the X-2C air vehicle. The Kroma 100i twin cylinder engine was chosen for its high powerto-weight ratio with 10 hp and a weight of only around 5 lb. It features an electronic ignition (EI) system powered by a 6 V battery and requires a mixture of gasoline and 2-stroke motor oil. The engine is placed in a pusher configuration using a Xoar tri-blade beechwood 24 in x 10 in pusher propeller. Proper air flow over the cooling fins is blocked due to the engine position behind the fuselage. To maintain an operational temperature during flight, a cowling was designed and fabricated to provide sufficient airflow for engine cooling. Airframe Testing Fuel-Consumption Tests X-2C is equipped with a 75 oz fuel tank that is located at the plane s center of gravity and uses regular unleaded fuel mixed with two-stroke oil. The tank weighs only three and a half pounds when full, providing a nice balance between flight time and overall weight. With this fuel capacity, X-2C has a flight time of approximately 1 hour at a cruise speed of roughly 40 knots. While at full throttle, 6300 rpm, Xawk uses three ounces per minute. This will easily allow for completion of the mission as Team X-ipiter will not be flying at full throttle since that exceeds the speed limitations of the competition. The use of regular unleaded fuel guarantees accessibility and availability no matter the location and a lower price compared to glow fuel. Flight Test Plan 1 The primary objective of the first flight test was to demonstrate flightworthiness, evaluate airframe stability and control, and allow time for the test pilot to familiarize himself with the aircraft. Flightworthiness was determined by balancing the location of the center of gravity and inspecting the structural integrity on the ground. Airframe stability numbers were confirmed by the pilot after takeoff and compared to prior calculated figures. Additional time for the test pilot was allocated, as he was new to the team. The flight test was performed with only the bare minimal components, including the R/C receiver, the DPSI Twin Maxi, and batteries. No imagery or autopilot software was tested, as those components were not present for the test. Three members of the team were required to ensure safe operation: the aircraft pilot, a spotter, and a safety technician. Prior to flight, range checks were performed again using the manufacturer s instructions. 16

17 After the personnel were briefed and checklist completed, the test was ready to begin. The pilot taxied the aircraft and performed takeoff and climb out. Once X-2C was cruising at an altitude around 300 ft, the pilot trimmed the controls while flying simple rectangular patterns. In accordance with the test plan, a few controlled approaches were flown before the actual landing attempt. The pilot landed X-2C successfully and gained much experience from the flight. Safety In accordance with the given requirements, the system payload is 25 lbs, and the airframe system component after fabrication is 30 lbs. This meets the weight requirements of the system. The Piccolo LT allows a manual override, allowing the safety pilot to take control of the airplane at any point. Also, X-2C uses three fail-safe modes: return home, flight termination, and a backup receiver. If the communications control signal to the autopilot is lost for 20 seconds, the airplane immediately sets a course to return to the predefined home waypoint. If the airplane does not regain a communication link and GPS after 20 seconds, the system will automatically terminate flight, as directed by the competition rules. In the event of a complete loss of the autopilot control, the DPSI will automatically switch to the backup receiver for manual control. The Kroma 100i twin-cylinder engine provides a level of reliability unmatched by any previous Xawk systems. Having two cylinders reduces the chance of an engine stall and emergency deadstick landing, an event that resulted in the crash of the X-2B UAS. Having two cylinders does, however, require a larger displacement based on what is commercially available. The 100 cc, 10 hp engine exceeds the power requirements and has a good power-to-weight ratio. The Kroma 100i twin cylinder engine features an EI system, allowing for the propulsion system to be started with a minimum number of prop cranks and resulting in a safer start procedure. To avoid starting the engine prior to the team being prepared, two safety switches have been installed. The first switch resides on the safety pilot s transmitter; the second switch was placed on the fuselage of the airframe. Both switches must be armed for power to be provided to the EI system. The airframe has a cruise speed of 45 knots. This speed allows for stable flight, successful use of the imagery system, and is clearly under the maximum airspeed specified by the competition SOW. The twisted and shielded wires used in X-2C are 22 AWG. The independent batteries for the GNC and Imagery subsystems use polarized Molex connectors. Separate power supplies prevent a failure in one subsystem to affect the other. Signals are passed through DB-9 and DB-25 connectors, and 6-amp toggle switches ensure reliable operation to withstand vibrations. Fasteners are composed of two divisions: fasteners for the payload and fasteners for the airframe. Payload fasteners consist of screws, nuts and bolts, dual-lock, industrial strength Velcro, and zip ties. Airframe fasteners include bolts with nuts, locking nuts, and t-nuts. Phase IV: System Integration, Testing, & Evaluation Payload Integration Figure 18 below shows the fuselage with integrated onboard Avionics components. The main considerations in placing the payload components are the aircraft center of gravity, engine vibrations, and external antennas. The components are moved as far forward as possible to help locate the center of gravity at the quarter-chord of the wing. Antennas are located externally to efficiently utilize internal space and to avoid signal interference with the carbon fiber in the hybrid composite fuselage skins. The camera is placed far from the engine to help reduce vibrations and improve image quality. The Video IP server and Ethernet Bridge are between the front bulkhead and the camera dome for weight balance and for location near the camera to reduce wiring lengths. The backup R/C receiver is mounted externally behind the camera, and the DPSI Twin Maxi is directly above inside the fuselage. The close distance between the receiver and DPSI is also effective in reducing the length of the wiring. 17

18 The Piccolo LT autopilot is mounted in the center fuselage section at the center of gravity of the aircraft. Due to the location of the fuel tank and main landing gear beneath the autopilot, the antenna is wired forward of the bulkhead and placed externally forward of the main gear. The rearmost compartment houses the electronic ignition module and throttle servo for the engine. Lead ballast weights are located in the fuselage near the nose cone and are easily adjusted to fine tune the location of the center of gravity. The ease of removal and adjustment of the lead weights is advantageous as the payload weight distribution changes in a series of test flights with different onboard components. The surveillance system s effectiveness was increased by applying a damping system. As with each of the other Xawk systems, the airframe and avionics systems encounter vibrations due to the powerful engine. In order to decrease the vibration effects on the payload, two damping systems were applied to the X-2C UAS: the engine was mounted using an isolation mount, and the surveillance system was affixed with a double vibration suppression mount using foam-rubber donuts between the camera and camera-mount and springs between the camera-mount and longerons. Figure 18. X-2C Fuselage with Onboard Avionics Integrated Testing X-ipiter has performed a series of tests to validate X-2C s performance and prepare to operate the system at the SUAS competition. All tests were designed with the foremost concern for safety of the system and the operators. Each testing procedure begins the day before the test with a detailed briefing involving the pilot, required personnel, and team advisors. The personnel go through a dry run of the test, and necessary changes are made and noted to the test plan. The necessary batteries are then charged, and required equipment is gathered for the airframe assembly and groundstation operation. Another briefing is given the day of the test to review the process and especially to note any changes. At the flight line, each person turns to the appropriate section of the test plan for the task. The team lead coordinates the personnel and completes the checklists during preparation of airframe and groundstation. The test proceeds when all personnel, team lead, and faculty advisors are ready. Flight Test Plan 2 The objective of the second flight test was to evaluate the performance of the backup R/C system with the imagery subsystem operating and to continue to familiarize the test pilot with the aircraft. The onboard components included the R/C receiver, DPSI Twin Maxi, Sony D70, Video IP server, Microhard Ethernet bridge, and batteries. Due to the autopilot being removed for this test, the XBS software was not tested with new pictures. The second flight test required four personnel to ensure safe operation. The four required roles were the aircraft pilot, spotter, safety technician, and XCCS operator. Before the flight, a range check for the R/C receiver was performed on the ground with the 2.4 GHz Microhard Ethernet bridge running. The FS button on the transmitter was held per manufacturer s instructions, and the demonstrated range was well over 50 paces. The pilot remained standing away from the aircraft while two personnel rotated the 18

19 aircraft to test for different headings as well as pitch and bank angles. Some interference occurred at certain orientations, and the power was turned down for the Microhard. Adjustments were made until the power was low enough to observe no interference with the R/C transmitter. With the power set at 200 mw, the ground range check was successful at all orientations. The preparation checklist was completed at the flight line, and the test was ready to begin. Flight proceeded with a successful takeoff and climb out, but interference was evident in the downwind and base legs of the pattern. The aircraft recovered and was landed immediately. The landing was successful, and the flight verified the operation of the imagery subsystem and presented the problem of interference. The solution was to move an antenna to the top of the plane to separate the conflicting signals. Flight Test Plan 3 The objective of the third flight test was to achieve full system integration, autopilot tuning, and test and evaluation. The autopilot subsystem was added to the existing aircraft system setup used in Flight 2 as well as incorporating XBS into the groundstation, seen below in Figure 19. In addition to the personnel from the previous flight, two more groundstation personnel were required to operate the autopilot command center and the XBS computers. Figure 19. Groundstation Transceivers, Antennas, and Transmitter Before the flight, a complete autopilot subsystem check was performed in addition to the R/C and camera system checks performed in the previous flight. The aircraft is rotated in all directions to verify the correct operation of the inertial measurement unit (IMU) and all onboard sensor values are checked or calibrated. All control surface deflections were verified by the pilot with the aircraft on manual control. Once manual control checks are completed, aircraft control is given to the autopilot and all control surfaces are checked again. After all ground checks and preflight checklist were completed, the pilot was given manual control and proceeded to takeoff. The aircraft was flown on the desired flight path while the autopilot command center operator tuned gains and model parameters until the aircraft performed as desired. The aircraft was landed and refueled several times over the course of the flight testing. Also, images were recorded during the flight and tested in the XBS software. Slight error was recognized in the XBS calculations from time delays between taking pictures and reading the Piccolo log file. 19

20 Phase V: Flight Demonstration The culmination of X-ipiter s design process is Phase V, the Flight Demonstration at the 2009 Student UAS Competition. The team has prepared using a systems engineering approach in the requirements, design, fabrication, testing, and integration of the components and subsystems. Full system tests have verified the flightworthiness of the airframe and performance of the GNC and Imagery subsystems. Before departing for competition on June 16, X-ipiter plans to perform more tests to assess the readiness of the X-2C Unmanned Aircraft System, seen below in Figure 20. Figure 20. X-2C Unmanned Aircraft System First, Flight Test 4 will perform autonomous takeoff and landing as an extension of Flight Test 3. All components will be onboard and thoroughly tested on the ground before proceeding to flight. The tests are expected to follow SIL and HIL flight simulations. Finally, Flight Test 5 will be a full practice mission, the objective of which would be to ensure the system and team is prepared for the flight demonstration portion of the competition. The system will be tested as a whole with all subsystems operational and all groundstation computers linked in real time. The flight time is limited to the system constraint of 40 minutes. The imagery subsystem software and operation will also be tested during this flight. Based off of targets from previous years of the competition, targets were designed and built for this flight test. These practice targets are shown below in Figure 21. Targets were designed to challenge the imagery subsystem in two ways: account for targets not properly identified in past competitions and combine similar colors on a single target. As with the competition, the targets spell out Raspet, representing the flight research facility at Mississippi State University and X-ipiter s home for all six years. The GNC, XCCS, and XBS operators will also gain experience in completing the mission during this practice mission. The XBS operator and assistant will need to identify all target parameters during the flight as will occur during the actual mission flight during competition. Figure 21. Design for Targets to be Used in Practice Mission 20