Autonomous Navigation and Landing Tasks for Fixed Wing Small Unmanned Aerial Vehicles

1、 Acta Polytechnica Hungarica Vol. 7, No. 1, 2010 Autonomous Navigation and Landing Tasks forFixed Wing Small Unmanned Aerial VehiclesSefer Kurnaz, Omer etinTurkish Air Force Academy, ASTIN Yesilyurt, Istanbul, 34807 T.tr, .trAbstract: Autonomous control of UAVs has

2、become a popular research topic in recentyears. This paper is concerned with the flight of UAVs (Unmanned Aerial Vehicles) andproposes fuzzy logic based autonomous flight and landing system controllers. Three fuzzylogic modules are developed under the main navigation control system and three more fo

3、rthe autonomous landing control system to control of the altitude, the speed, and theposition against the runway, through which the global position (latitude-longitude) of theair vehicle is controlled. A SID (Standard Instrument Departure) and TACAN (Tactical AirNavigation) approach is used and the

4、performance of the fuzzy-based controllers isevaluated with time based diagrams under MATLABs standard configuration and theAerosim Aeronautical Simulation Block Set which provides a complete set of tools for rapiddevelopment of detailed 6 degree-of-freedom nonlinear generic manned/unmanned aerialve

5、hicle models. The Aerosonde UAV model is used in the simulations in order todemonstrate the performance and the potential of the controllers. Additionally, some visualtools are deployed in order to get visual outputs that aid the designer in the evaluation ofthe controllers. Despite the simple desig

6、n procedure, the simulated test flights indicate thecapability of the approach in achieving the desired performance.Keywords: UAV, Fuzzy Logic Controller, Autonomous Flight Control, AutonomousLanding System1 Introduction In the literature, it can be seen that the research interests in control and na

7、vigationof UAVs has increased tremendously in recent years. This may be due to the factthat UAVs increasingly find their way into military and law enforcementapplications (e.g., reconnaissance, remote delivery of urgent equipment/material,resource assessment, environmental monitoring, battlefield mo

8、nitoring, ordnancedelivery, etc.). This trend will continue in the future, as UAVs are poised toreplace the human-in-the-loop during dangerous missions. Civilian applications ofUAVs are also envisioned such as crop dusting, geological surveying, search andrescue operations, etc. 87 S. Kurnaz et al.A

9、utonomous Navigation and Landing Tasks for Fixed Wing Small Unmanned Aerial Vehicles One of the important endeavors in UAV related research is the completion of amission completely autonomously, i.e. to fly without human support from take offto land on. For unmanned aerial vehicle systems to achieve

10、 full autonomy, smarterairplanes need to be developed. Full autonomy means performing takeoffs,autonomous waypoint navigation and, especially landings while the craft ishardest to control under computer control autonomously. The ground stationcontrol operator plans the mission and the target destina

11、tion for reconnaissanceand surveillance. UAV then takes off, reaches the target destination, completes thesurveillance mission, and turns back to the base and lands on autonomously.Navigation is the topic which is most studied about. In literature, many differentapproaches can be seen related to the

12、 autonomous control of UAVs; some of thetechniques proposed include fuzzy control 1, adaptive control 2, 3, neuralnetworks 4, 5, genetic algorithms 6. Capabilities of autopilot systems areimportant to successfully complete the mission of an UAV. A number of differentautonomous capabilities may be re

13、quired to be exhibited during a flight, likeautonomous take off, navigation and autonomous landing.In Section 2 of the paper the design of the navigation system with fuzzycontrollers used for the autonomous control of the UAV is described and theautonomous landing model is defined by using the landi

14、ng parameters given.Section 3 starts with the definition of the basic flight pattern for a UAV and thenexplain a sample mission plan which includes SID (Standard InstrumentDeparture) and TACAN (Tactical Air Navigation) procedures. A number ofsimulation studies are carried out with the fuzzy logic ba

15、sed autonomousnavigation and landing system and some typical results are presented in Section 4.Finally the concluding remarks and some plans about future work are given inSection 5.2 Fuzzy Logic-based System Design As shown in Fig. 1, there basically are two separate computers used in anautonomous

16、UAV. One of them is the flight computer and the other is the missioncomputer. UAV flight computer basically sets the control surfaces to the desiredpositions by managing the servo controllers in the defined flight envelope suppliedby the UAV navigation computer as a command, reads sensors and commun

17、icateswith the mission computer and also checks the other systems in the UAV (enginesystems, cooling systems, etc.). The navigation computer and the landing systemare a part of the mission computer. The mission computer carries out many otherduties beside navigation, like payload control, communicat

18、ion with GCS, etc., thenavigation computer is used for flying over a pattern which is designed before theflight or while flying in LOS (Line Of Sight). When GCS (Ground ControlStation) gets control of UAV, the navigation computer goes into a passive state.During autonomous flight, the navigation com

19、puter gets the position values from 88 Acta Polytechnica Hungarica Vol. 7, No. 1, 2010 sensors (GPS receiver, altimeter, INS (Internal navigation System), etc.) and thenmatches these values (the current position) with the desired position values (thewaypoint values). The navigation computer then det

20、ermines the requiredmaneuvers of the UAV to reach the goal position and sends these data to the flightcomputer to apply them to control surfaces.Figure 1 UAVs electronics architecture Landing is one of the most critical parts of a flight, because, like in traditionalaircrafts, UAVs aim to land at mi

21、nimum air speed, consequently the stabilityconditions are severe and the maneuvering abilities are limited. A total of 65Predators have crashed to date and thirty-six of the crashes were attributed tohuman error, and half of those occurred during landing 7. The difficulties intaking off and landing

22、arise mostly from instinctual factors, because pilots usetheir feelings in these periods of flight, such as feeling the ground rush and havingperipheral vision. For a kite pilot, landing is a process which aims to see the wingsinside of runway, but when a pilot is not in the cockpit, such feelings d

23、o not existany longer. Because of these reasons, the manual control of an UAV from theground is not a good alternative in the case of an emergency, especially duringtakeoff and landing.The architecture used by the authors in 1 for their work on a Fuzzy Logic BasedNavigation Control System (FLBNCS) f

24、orms the basis of the architecture for theFuzzy Logic Based Autonomous Landing System (FLBANS). The missioncomputer has two main parts in this work, FLBNCS and FLBALS, as shown inFig. 2. After getting the sensor values from the sensor interface, both FLBNCSand FLBANS calculate the desired attitude o

25、f the UAV attitudes which must beachieved by the flight computer. Then flight computer selects the correct 89 S. Kurnaz et al.Autonomous Navigation and Landing Tasks for Fixed Wing Small Unmanned Aerial Vehicles commands between the navigation c and the landing system commands. If UAV isin the final

26、 approach pattern it uses the landing systems commands, else it uses thenavigation computer commands as inputs. The flight computer then calculates thecontrol surfaces and the throttle positions by using its direct sensor inputs and thecommand inputs to reach the desired attitudes. The flow of this

27、process can beseen in Fig. 2.Figure 2 Autonomous System Design The operation of the navigation computer proposed in this paper is fuzzy logicbased. Basically, a fuzzy logic system consists of three main parts: the fuzzifier,the fuzzy inference engine and the defuzzifier. The fuzzifier maps a crisp i

28、nputinto some fuzzy sets. The fuzzy inference engine uses fuzzy IF-THEN rules froma rule base to reason for the fuzzy output. The output in fuzzy terms is convertedback to a crisp value by the defuzzifier.In this paper, Mamdani-type fuzzy rules are used to synthesize the fuzzy logiccontrollers, whic

29、h adopt the following fuzzy IF-THEN rules:R)AND.AND(x isX n)THENy isY1 ,., y isYk(i) : If (x isX1i i i i1 n 1 k(1)Rl is the lth rule x = (x ,.,xn)TU and y = (y ,., yn)TV arewhereiithe input and output state linguistic variables of the controller respectively,U,V n are the universe of discourse of th

30、e input and output variables( ,., x ) U and (y ,., yk) V are the labels in linguisticrespectively, x1T Tk 1terms of input and output fuzzy sets, and n and k are the numbers of input andoutput states respectively. 90 Acta Polytechnica Hungarica Vol. 7, No. 1, 2010 We consider a multi-input and single

31、-output (MISO) fuzzy logic controller(k =1) , which has singleton fuzzifier. Using triangular membership function,algebraic product for logical AND operation, product-sum inference and Centroiddefuzzification method, the output of the fuzzy controller has the following form: NM(i=1xil(x )yiiy =jl=1(

32、2) NMxi (xi)ll=1i=1Where N and M represent the number of input variables and total number of rulesrespectively. xil denote the membership function of the lth input fuzzy set forith input variable.theIf the fuzzy controller types in literature are reviewed, it can be seen that there aretwo main class

33、es of fuzzy controllers: one is position-type fuzzy controller whichgenerates control input (u) from error (e) and error rate(e) , and the other isvelocity-type fuzzy logic controller. The former is called PD Fuzzy LogicController and the latter is called PI Fuzzy Logic Controller according to thech

34、aracteristics of information that they process and system has two inputs, theerrore(t) and change of error e(t) , which are defined bye(t) = y yref(3)e(t) = e(t) e(t 1)(4)Where yref and y denote the applied set point input and plant output,respectively. The output of the Fuzzy Logic Controller is th

35、e incremental changein the control signal u(t) . PD type fuzzy logic controller can be seen in Fig. 3.Then, the control signal is obtained byu(t) = u(t 1) + u(t)(5)Figure 3 PD Type Fuzzy Logic Controller 91 S. Kurnaz et al.Autonomous Navigation and Landing Tasks for Fixed Wing Small Unmanned Aerial

36、Vehicles As the second system in the architecture, FLBALS system uses the position inputsto calculate the exact location against the runway. It determines the error andcalculates the corrective maneuvers by using three additional fuzzy logicsubsystem blocks (in addition to the three blocks used for

37、navigation). First fuzzyblock is the lateral fuzzy logic controller which resolves the lateral errors inFLBALS. The second block is the vertical fuzzy logic controller which resolvesthe altitude errors and the last one is the speed fuzzy logic controller which tries toachieve the desired speed for t

38、he current conditions in FLBALS.Figure 4 Fuzzy Logic Based Autonomous Landing System Design The inputs to these fuzzy logic blocks are provided by different systems likeILS/INS and GPS 8, 9, laser based systems 10 or by vision based algorithms11. Other inputs of these blocks are the landing pattern

39、flight plan or the manualcommands issued by the Ground Control Station (GCS). The inputs of the fuzzyblocks can be seen in Fig. 4 and the surface diagrams of the blocks can be seen inFig. 5. 92 Acta Polytechnica Hungarica Vol. 7, No. 1, 2010 3 Autonomous Flight Model The test pattern used in this st

40、udy includes a box pattern of Yalova Airport(LTBP) 18-26 Runway to show that UAV can fly autonomously a pattern which isdesigned for aircrafts if it has enough performance parameters. In classical SIDand TACAN maps, a waypoint is defined with the radial angle and the distancebetween the VOR (VHF Omn

41、i-directional Radio Range) station and the waypoint.After transformation of the waypoints as GPS coordinates, UAV can apply SIDdeparture and TACAN approach as a mission plan without a VOR receiver. InFig. 6, the top and the side views of the test flight pattern is shown. This is a kindof box pattern

42、. There are some important points which must be defined as GPScoordinates, like the initial approach point (IAP), the last turn point (LTP), the lastapproach point (LAP), the minimum altitude point (MIN) and the downwind turnpoint (DWTP). The UAV must reach the minimum altitude before the MIN pointa

43、fter takeoff. Then the UAV continues to the MIN point and starts to turn to reachDWTP. The particular set of these points that is used in the simulation studies isshown in Table 2. Each point of the pattern is represented by three values, thelatitude and the longitude as the GPS position and the alt

44、itude as the verticalposition.a) Lateral Fuzzy Control Surface b) Speed Fuzzy Control Surface c) Vertical Fuzzy Control Surface Figure 5 Control Surfaces In this study, the UAV is considered to take off manually. Autonomous navigationstarts when the UAV reaches the MIN point. It then reaches the way

45、 points inorder and finishes when the UAV reaches the REP. The definition of each pointincludes speed, altitude and position (longitude and longitude coordinates) values. 93 S. Kurnaz et al.Autonomous Navigation and Landing Tasks for Fixed Wing Small Unmanned Aerial Vehicles Figure 6 The test patter

46、n of the UAV autonomous flight To land, the aircraft must reach to the IAP and then it aims to reach the LTP andthe LAP in order. After reaching the LAP, if airfield is not suitable for landing, itgoes into a holding pattern. When the airfield becomes ready to land, the UAVcompletes the turn until t

47、he LAP is reached and goes into the final approach stage.Table 1 Definitions of test pattern waypoints. Point NameCoordinate (GPS)Altitude (feet)Runway Starting Point (RSP)Minimum Altitude Point (MIN)Down Wind Turn Point (DWTP)Initial Approach Point (IAP)Last Turn Point (LTP)N40 41 39.20 E29 22 34.8

48、1N40 49 56.67 E29 22 34.81N40 49 56.67 E28 18 43.00N40 41 50.00 E29 17 50.00N40 32 26.55 E29 18 43.00N40 32 26.55 E29 22 34.81N40 40 56.54 E29 22 34.816150015001700150012006Last Approach Point (LAP)Runway End Point (REP)The values in Table 1 are contour a huge left side box pattern for Yalova (LTBP)

49、Airfield 18-36 runway. First turn point in pattern is MIN point and it is 9 nm awayfrom runway. Downwind leg is 3 nm away from runway and parallel to the 18-36runways. Also LAP is 9 nm away from runway too. All the altitude values aremean sea level (MSL) in table. 94 Acta Polytechnica Hungarica Vol.

50、 7, No. 1, 2010 To accomplish a successful landing, there are three main attributes which must beunder control. First of them is the lateral position of the UAV with reference to therunway. As has already been stated, the goal is to touchdown on the lateral middlepoint of the runway like in Fig. 7.

51、The second attribute is the vertical position,which is the AGL (above ground level) altitude of the UAV. It is a dynamic valuesince it changes according to the distance to the runway, but the usual glide pathangle is 3 degree in aviation literature as in Fig. 7. The glide path angle is 3degrees in n

52、early all the airfields in the world if there is no obstacle in this 3o path.The last main attribute is the speed. The speed value is a static value and itdepends on the aircraft characteristics. The main aim is keep the desired speedvalue during the period of the final approach.Figure 7 Desired Dow

53、nward Velocity In order to obtain the lateral position of the UAV with reference to the runway,different techniques can be used, like image processing 11 or radio basedposition calculators 8 and ILS (instrument landing systems) 9. To measure thealtitude and the speed of the UAV, laser altimeters and pito systems can be usedrespectively 10. In t