Sei sulla pagina 1di 17

Low-Cost Autonomous Civilian Unmanned Aerial Systems

by Christian Stewart March 30, 2012

Biology 9 Period 4 Ms. Pi Harvard-Westlake School Los Angeles, CA

Abstract Unmanned aircraft, commonly called Unmanned Aerial Vehicles (UAVs), when combined with ground control stations and data links, form Unmanned Aerial Systems (UASs). UASs pose a wide variety of applications for civilian use, including imaging large areas for survey. I researched current UASs in use, and found no aordable solutions suitable for civilian drone applications. The objective of the project was to design and develop a low-cost Autonomous Unmanned Aerial System (AUAS) capable of meeting the needs of civilian applications. The system was designed specically for photographing interest areas for survey via a small xed wing Unmanned Aerial Vehicle. A list of requirements for the UAS was made. A design was created to conform to these requirements, and the components were acquired with some elements donated by their suppliers for development. The system was built to follow the design, and user-friendly Mission Control software was developed to allow real-time control and monitoring of the UAV (Unmanned Aerial Vehicle). An experiment to verify the systems capability to execute the sample application, Aerial Imaging for Survey, was devised. Interest areas were selected by an operator with the Mission Control software, and the drone was deployed under full autonomous control. Data measuring every aspect of the planes ight, most importantly locational, airspeed, and altitude was electronically recorded. This test procedure was repeated with ve diering interest areas. Test ights were executed at the Apollo XI model aircraft eld in Los Angeles on Febuary 13, 18, and 19. The data from each test was analysed by calculating the area imaged relative to the UAV, and then comparing this area to the original interest area for the ight. This information was used to calculate the percent interest area imaged for each test. If a percent area imaged was greater than 100%, it was rounded down to 100%. The average percent area imaged was calculated to be 100%. The total cost of the system if deployed commercially ($594.36) was compared to the published cost of a comparable drone (Ocatrons SkySeer) in commercial use ($25,000) and was found to cost $24,405 less than current solutions. The results of the experiment were analysed. It was concluded that the nal design was capable of meeting the needs of a common civilian application, and was extremely low-cost in comparison with other systems currently in use.

Introduction

According to an article published by the Washington Times, February 7th 2012, the Federal Aviation Agency projects that 30,000 unmanned drones will be in the United States skies by 2020. [12] Unmanned aircraft, commonly called Unmanned Aerial Vehicles (UAVs), when combined with ground control stations and data links, form Unmanned Aerial Systems (UASs). UASs pose a wide variety of applications for civilian use;[16] they can take the place of manned vehicles in a variety of situations from scientic atmosphere studies to imaging terrain for survey. [13, 8, 7, 11, 10] The majority of Unmanned Aerial Systems (UASs) in use today were developed for large scale military and government applications. National spending on UASs has increased from $284 million in 2000 to $3.3 billion in 2010.[13, 14] These drones are extremely expensive, dicult to build and maintain, and are not a viable solution for civilian use. A report published by NASA[13] explores the capability and applications of civil UASs, and documents potential future civil missions for UASs based on user-dened needs. Applications explored by the report range from land management to Earth science. One of the highest priority needs for civil UASs mentioned in the report was autonomous mission management, as well as quick deployment and high availability. Today, current UASs cannot oer the required elements of civil applications, as they are very expensive, ranging from $25,000[9] to $30,300,000[4], and a majority of these drones do not oer autonomous control. The goal of this project was to design and develop a low-cost Autonomous Unmanned Aerial System (AUAS) capable of meeting the needs of civilian applications.

II

Materials and Methods

Aerial Imaging for Survey was selected as a sample application for the objective design of the system. A list of requirements for the AUAS (Autonomous Unmanned Aerial System) was made. A design was created to conform to these requirements, and the required materials were acquired with some elements donated by their suppliers for development.

Aerial Vehicle Design

The aerial system was developed rst. The target airframe was of a xed-wing foam aircraft. After considering various airframes and electronics, a table of the nal equipment list was made, including each components commercial cost. The nal cost of the system was determined by calculating the sum of the commercial costs of each of the components, and was noted in a data table including the researched prices of comparable drones in use. Name Bixler Airframe HXT900 Servo x4 Thunderbird-36 Purpose Base Airframe Control Servos Electric Speed Control Propulsion Motor Computer Control Vendor HobbyKing HobbyKing Castle ations Emax uThere CreCost $41.44 $8.97 $49.95 $25.95 Comment EPO Foam, Fixed Wing, 1400mm, Powered Glider Servos to move control surfaces. Speed control for motor.

2836-24 Brushless Inrunner uThere Ruby w/Sensors

Main motor to propel plane.

5.8ghz 200mw TX/RX

RMRC-PICO Camera 1300mah Turnigy Nano- Power Tech Li-Po Battery Source

Video Data-link 6-24v Camera

$345.00 Computer Controller to execute ground station commands. HobbyKing $64.90 Video transmitter and receiver for live video feed. ReadyMadeRC $48.99 Onboard camera for live video feed. HobbyKing $9.16 Battery to provide power to onboard electronics.

Figure 1: Airborne Equipment and Commercial Costs

Ground Station

Next, a ground station capable of real-time control and monitoring of the aircrafts mission was designed. The ground station consisted of a Viore 8 portable TV for visualizing the live video feed, as well as two USB connections to a laptop running custom Mission Control software, which was based on the .NET Framework 4 for an interactive map display. Code was written to provide a easy to use interface for mission control, waypoint designation, interest area designation, and monitoring of essential telemetry received from the aerial system.

Figure 2: Screenshot of the Custom Mission Control software..

Experiment

An experiment to verify the systems capability to execute its mission was devised. Interest areas were input by an operator to the Mission Control software, and the aerial system was deployed under full autonomous control. Once airborne, the system was commanded to begin fully autonomous execution of imaging the interest area using the Mission Control software running on the laptop. The interest area imaging mission was congured to y at a constant 350 feet. Once the imaging was complete, the system was commanded to execute a fully autonomous landing. After each ight, data les containing a history of every detail observed by the airborne control system were copied from a Mini-SD card removed from the controller. The test procedure was repeated with ve diering interest areas, during ve test ights on Febuary 13, 18, and 19 at the Apollo XI Model Aircraft Field in Los Angeles, California.

Figure 3: Site map of the Apollo XI Model Aircraft Field (image credit Google Maps)

Data Processing

Data corresponding to position and altitude was extracted from the data les and exported from uTheres Ground Control (unreleased) and imported into the projects custom software, where a computer algorithm calculated the area imaged. recorded locations = locations from data le area imaged = empty array of locations for all location in recorded locations do cameraview = orthographic circle (from top-down view) cameraview radius camera altitude area imaged += co-ordinates in imaged area end for
Figure 4: Area imaged algorithm outline.

The algorithm calculated the area imaged by iterating over each stored data point in order. For each stored data point, a top-down orthographic circle representing the view of the downward facing camera was calculated. A geometric proof was created to determine the radius of the cameras view, which was visualized by creating a horizontal 2D triangle.

Figure 5: Right triangle representing camera geometric proof.

The top angle of the triangle had a measure of 90 degrees, as the cameras eld of view is 90 degrees. The geometric altitude was drawn from the 90vertex to the hypotenuse of the triangle, producing two right triangles. The horizontal leg of the resulting right triangle 5

represented the radius of the observed circle, while the vertical leg represented the altitude of the camera above ground. The two non-right angles were concluded to measure 45, as the angle representing the 90eld of view of the downward-facing camera was bisected to form two 45angles, and the third must also measure 45as the sum of the measures of the angles in a triangle is 180, and 180 90 45 = 45. The triangle was therefore a 45, 45, 90special right triangle, and thus its vertical leg was equal to the horizontal leg, and the hypotenuse was equal to the measure of either of the legs times the square root of two. This proof was applied to the known altitude and eld of view and altitude of the camera. Let altitude = x and the radius of observed circle = y. Due to the properties of the 45-45-90triangle, it was concluded that x = y, and thus the altitude was equal to the radius of the observed circle. This calculation was used to determine the area imaged during the calculated data point using the formula A = r2 . The resulting area measurement was applied to the positional data for that data point to determine the GPS coordinates of the imaged area to an accuracy of 0.001 degrees, or about 5.8 feet of terrain. This was repeated for each of the recorded data points. After processing all of the recorded data with the imaged area algorithm, the software removed co-ordinates from the imaged area array that were not a part of the original interest area (and thus eliminated the extra area imaged from the % image area calculation) and compared the co-ordinates of the original interest area with the co-ordinates of the imaged area. The percent interest area imaged was calculated using the formula % Interest Area Imaged = #imaged interest points/#total interest area points. This calculation procedure was repeated for each of the test ights, and the results recorded in a table.

III

Results

The system was consistently able to execute its mission, imaging more than 100% of the interest area during every test. From this information, we can deduce that the design was a success, capable of fullling its target application. Flight # Percent Interest Area Imaged 1 100% 2 100% 3 100% 4 100% 5 100% Date 2/13/2012 2/18/2012 2/19/2012 2/19/2012 2/19/2012

Figure 6: Percent Interest Area Imaged During Experiment

The most comparable UAS currently in use is the Ocatron SkySeer. When comparing the specications of the SkySeer to the specications of the proposed system, it can be seen that the two systems are very similar. Feature Ocatron Sky-Seer Proposed System GPS Navigation Yes Yes GPS Control Yes No Endurance (Mins) 50 minutes 30 minutes Weight (pounds) 4 1.5 Speed (knots) 21 21 Range (miles) 2 3.1 Pan (degrees) 160 140 Tilt (degrees) 90 45 Autonomous Flight Yes Yes Price (USD) $25,000 $594.36
Figure 7: Feature comparison of comparable reference system and proposed system.

Both are able to achieve autonomous control. According to fact sheets published by Ocatron[9], the extent of autonomous ight in the SkySeer is to waypoint specication and ight. Although the proposed system is capable of this, it is also capable of generating these paths on its own, and modifying its ight plan in real time. This poses an advantage as often conditions such as wind can vary the path of a small plane, which would require minute 7

changes to the ight plan to eectively image the target area. Both systems are capable of greater than two mile ight range, as well as pan-tilt movement of the camera. Although both utilize GPS navigation, the SkySeer is also capable of GPS control links, something that is yet to be implemented in the proposed system. Endurance was greater for the SkySeer than the proposed system in testing, however a larger battery could easily extend the ight time to more than an hour. System Name Vendor Name Price Point Proposed System Christian Stewart $594.36 SkySeer Ocatron $25,000.00 Wasp Aerovironment $49,000.00 Raven Aerovironment $173,000.00 Predator General Atomics $4,500,000.00 Grey Eagle General Atomics $8,000,000.00 Reaper General Atomics $30,300,000.00
Figure 8: Commercial Price Point of Proposed UAS vs. Estimated/Published Price Points of UASs in Use

The primary dierence between the two systems is their price. The Ocatron SkySeer costs between $25,000[5] and $30,000[15], while the proposed system costs only $594.36. The total cost dierence is approximately $25,405.64. The results of the experiment prove that the proposed system is capable of executing civil UAS applications, while fullling two of the high priority needs stated in NASAs civil UAS capability report: high availability, at a low price point compared to current solutions, as well as autonomous mission management.

Figure 9: Chart representing price points of unmanned aerial systems currently in use.

IV

Discussion

The goal of the project was met; my low-cost Autonomous Unmanned Aerial System was capable of meeting the needs of civilian applications. As drones become more common and widely used, it is vital that work be done to extend them to the civil eld. In order for these drones to become practical for civilian applications, the needs outlined in NASAs civil UAS capability document[13] must be met, including high availability made possible by low costs, and autonomous mission management for regularity and ease of use. My experiment could have been improved by simulating a real-life scenario such as imaging a crop to record important information about its condition. Furthermore, more varied weather conditions could have been used to test the capability of the systems auto path correct and replanning. Engineering the prototype system in collaboration with a potential end-user would allow me to nd crucial elements required by the user and implement them. There are many other hurdles that must be passed before civilian UASs are a reality, for example, regulations regarding UASs limit their ight ceiling to 400 feet, a altitude which many UAVs must surpass in order to be eective. Further controversy has arisen regarding what is legal for these systems to image, and where they are allowed to y. For example, UASs have become increasingly prominent in the lm industry, however complaints have been made when private property has been lmed intentionally or not. Working on this project has lead me to wonder about and research more potential applications for low-cost AUASs. These systems can be applied to many other needs such as crop dusting, animal tracking, and disaster response. The existence of these highly available solutions would drastically improve the eciency and reduce issues faced by re ghters facing a wild re, or humanitarian aid attempting to help those in distress. I hope to see these systems applied in civil elds in the future, as I see great potential in their assistance of mankind. This science fair project has inspired me to plan many more revisions and applications for my system, and to continue development of low-cost UASs in the future. I plan to apply 10

the theory of low-cost autonomous drones to more durable airframes capable of harsher environments such as the heat generated by a re or a volcano, as well as those capable of carrying the weight of heavy scientic equipment. I also plan to increase the range and ight time of these systems, continually developing them to be more t for many civil applications. UASs are steadily becoming more prevalent in our world. I envision a future in which UASs are commonplace as a major piece of civilian activity, and plan to continue developing this idea to help usher in a world in which UASs are not just a matter of science ction or national security, but a major part in our lives as helpful eyes in the sky.

References Cited

References
[1] Bowes, Peter. High Hopes for Drone in LA Skies. BBC News. BBC, 06 June 2006. Web. 25 Mar. 2012. <http://news.bbc.co.uk/2/hi/americas/5051142.stm>. [2] Fernandez, Juan. Drones Now Being Used To Sell Pricey Southern Cali31 Oct. 2011. Web. 19

fornia Real Estate. CBS News. CBS Interactive,

Mar. 2012. <http://losangeles.cbslocal.com/2011/10/31/drones-now-being-used-to-sellpricey-southern-california-real-estate/>. [3] Gertler, Wired.com. Jeremiah. Conde Almost Nast 1 In 09 3 U.S. Jan. Warplanes 2012. Web. Is 20 a Mar. Robot. 2012.

Digital,

<http://www.wired.com/dangerroom/2012/01/drone-report/>. [4] Johnson, Robert. FAA: Look For 30,000 Drones To Fill American Skies By The End Of The Decade. Buisiness Insider. Buisiness Insider, 08 Feb. 2012. Web. 25 Mar. 2012. <http://articles.businessinsider.com/2012-02-08/news/31036604 1 dronesunmanned-aircraft-new-bill >.

11

[5] Mortimer, Gary. Los Angeles Sheris UAV Runs Headrst into the FAA. SUAS News, 22 June 2006. Web. 19 Mar. 2012. <http://www.suasnews.com/2006/06/12963/losangeles-sheris-uav-runs-headrst-into-the-faa/>. [6] Mortimer, Gary. Pygmy Rabbits Landscape. SUAS News. SUAS News, 21 Feb. 2012. Web. 19 Mar. 2012. <http://www.suasnews.com/2012/02/12181/pygmy-rabbitslandscape/>. [7] MQ-1B PREDATOR. Af.mil. United States Air Force, 5 Jan. 2012. Web. 19 Mar. 2012. <http://www.af.mil/information/factsheets/factsheet.asp?fsID=122>. [8] MQ-9 REAPER. Af.mil. United States Air Force, 5 Jan. 2012. Web. 19 Mar. 2012. <http://www.af.mil/information/factsheets/factsheet.asp?fsID=6405>. [9] SkySeer. Octatron. Ocatron. Web. 20 Mar. 2012.

<http://www.octatron.com/prodSkySeer.php>. [10] Pteryx UAV. Trigger.pl. Trigger Composites. Web. 20 Mar. 2012.

<http://www.trigger.pl/pteryx/Pteryx-UAV.php>. [11] RQ-11B RAVEN. Af.mil. United States Air Force, 14 Sept. 2011. Web. 19 Mar. 2012. <http://www.af.mil/information/factsheets/factsheet.asp?fsID=10446>. [12] The Skys Eyes: Remote Sensing in Archaeology. PBS. Public Broadcasting Service. Web. 20 Mar. 2012. <http://www.pbs.org/wgbh/nova/ubar/tools/>. [13] SkySeer UAV Drone Soars Over L.A. Inventions and Ideas from Science Fiction Books and Movies. Technovelgy, 15 June 2006. Web. 25 Mar. 2012. <http://www.technovelgy.com/ct/Science-Fiction-News.asp?NewsNum=654>. [14] UAC Commercial Applications Overview. uavm.com. UAV MarketSpace. Web. 20 Mar. 2012. <http://www.uavm.com/uavapplications.html>.

12

[15] United Timothy Warner.

States. H.

National

Aeronautics J. and

and Nagy,

Space Mark Agency.

Agency. A.

Nasa.gov. and Mar.

By Ryan 2012.

Cox,

Christopher Aeronautics

Skoog, 24

National

Space

Web.

<http://www.nasa.gov/centers/dryden/pdf/111760main UAV Assessment Report Overview.pdf>. [16] WASP III. Af.mil. United States Air Force, 14 Sept. 2011. Web. 19 Mar. 2012. <http://www.af.mil/information/factsheets/factsheet.asp?id=10469>. [17] United Room. By States. Jeremiah Congress. Gertler. Congressional Wired, 3 Jan. Research 2012. Web. Service. 25 Mar. Danger 2012.

<http://www.wired.com/dangerroom/2012/01/drone-report/>.

13

VI

Appendices

Figure 10: UAV in ight.

14

Figure 11: Retrieving data from drone.

15

Potrebbero piacerti anche