Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing

Transcription

1 Indian Journal of Science and Technology, Vol 9(46), DOI: /ijst/2016/v9i46/107504, December 2016 ISSN (Print) : ISSN (Online) : Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing Nikolay V. Kim *, Nikolay E. Bodunkov, Ivan G. Krylov and Elena A. Fedyaeva Federal State Higher Military Educational Establishment Nakhimov Black Sea Higher Naval School, Sevastopol, Russia; nkim2011@list.ru, boduncov63@hotmail.com, chwwmu@yandex.ru, fedyaeva.alena@yandex.ru Abstract Objectives: Automatic landing of an aircraft-type Unmanned Aerial Vehicle (UAV) on the ship is a complex research and engineering problem. In the general case, a set of problems must be solved subject to the initial position of the UAV and the ship, weather, visibility, reliable radio communication etc. in preparation for landing and capturing the initial point of a glide path for landing. These problems include the UAV flying to the ship location, searching for and identifying the ship, determining rolling parameters and estimating the position of landing markers on the ship deck. Method: Methods of UAV guidance to the ship search area have been compared. Conditions for the effective application of various methods of UAV guidance to the ship have been determined. Findings: It was demonstrated that insufficient attention is paid in available research to scenarios, when the ship may change its direction and a communication channel with the UAV is not stable due to the weather. The following tasks are discussed: determining conditions of physical realizability of the successful ship search; selecting a method to guide the UAV towards the ship, with view to ship maneuverability intensity and duration of transmission breaks to the UAV about current ship parameters according to the flight endurance minimization criteria; and determining UAV flight endurance to the ship location subject to the duration of transmission breaks. Conditions of the successful ship search have been determined and the ratio of the initial search area and UAV flight endurance has been shown. Improvements: The results of the study can be used in the development of autopilots for autonomous UAVs based on the ships. Keywords: Embarkment, Flight Endurance, Flight Planning, Unmanned Aerial Vehicle 1. Introduction Automatic landing of an aircraft-type unmanned aerial vehicle (UAV) on the ship is a complex research and engineering problem. Issues related to UAV landing onto a platform that does not have any special radio equipment are considered in 1. Required accuracy of implementation of the glide path at different altitudes is discussed. It has been shown that satellite navigation (that does not involve any auxiliary measures) does not ensure the required accuracy of UAV coordinate estimation for safe landing. Visual navigation methods are recommended in such cases, while referencing to ground-based visual references, and special markers in particular. Image processing algorithms that allow for solving visual navigation tasks to the required degree of accuracy are described. UAV embarkment arrangements are even more complex. Let us consider some aspects of implementation of this process. Overall, a set of problems must be solved subject to the given conditions, in particular the initial position of the UAV and the ship, weather, visibility, reliable radio communication etc. in preparation for landing and capturing the initial point of a glide path for landing: Flying the UAV to the ship location. It is assumed that we know ship coordinates or location area at the beginning of this process. Searching for and identifying the ship. This search implies scanning the search area to establish an * Author for correspondence

2 Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing energetic (in particular visual) contact of the scanner (UAV) with the object of search (the ship). Identifying landing references (markers), estimating their position, and determining rolling parameters. For example, embarkment that involves rolling is discussed in the research 2. Calculating necessary landing parameters and capturing a glide path for landing. A flight to the ship location area may take long. At the same time, route planning relies on the a priori information. If the initial distance between the UAV and the ship exceeds greatly maximum range of visibility of the ship by the UAV surveillance system (SS), but there is continuous reliable radio communication between the UAV and the ship available, ship coordinates may be transmitted regularly to the UAV, in which case planning and flying the UAV to the ship is implemented, using various methods according to the given efficiency criteria. In particular, when the ship course is known, the UAV may be directed to an estimated meeting point. For varying ship courses, however with regular information about ship coordinates available, various guidance methods may be utilized. For example, an approach to improve the proportional guidance law for landing onto the moving ship is discussed in 3. Building of a navigation system, which is necessary to embark the UAV in unfavorable conditions, is considered in 4. Development of a fully automated system to land the UAV on ships in the sea is examined in 5. A great share of research is devoted to arrange UAV flights to objects of search. Various search scenarios and corresponding control strategies are explored in 6. Methods of guidance and chases of moving objects against different initial conditions and with view to velocity-distribution laws and heading angles of the objects of search are considered. At the same time, it is assumed that targets of the scanner and object of search are opposite. A method of flight route planning for an UAV, considering flying around dangerous areas is discussed in 7. A minimum path-length algorithm for a maneuvering UAV is given in 8. Path-length algorithms for an UAV are also considered in 9. Studies to optimize UAV flight path, prevent any crashes, and enhance flight safety are presented in 10,11. Issues of UAV guided flight are dealt with in 12. Searching for the ship. Under special circumstances, e.g. when communication with the ship is lost, the necessity of additional ship search arises 6,13,14. Direct contact with the object of search is established by the detection procedure. The need of search arises, when there is not any visual contact with the ship and its coordinates are unknown (non-determined), in particular, when there is not any communication with the ship or the communications channel is unstable. If ship coordinates were known, but communication with the ship was lost for some time, and ship path is not known, a so called secondary search must be applied, in which case the area of possible search is extended against the undetermined path and depends on the increased time of lost communication and the possibility to clarify current ship coordinates. Ship identification is required, for example, when false objects may be available in the search area or range of visibility of the UAV SS. Detection of landing references and estimation of their position on the ship deck are necessary (with view to ship rolling parameters) to calculate the glide path for landing. Overall, search and identification of the object of search and detection of landing references (signs, symbols, or markers) are done by marking out a set of attributes In particular, various correlation algorithms that may be used to identify and detect reference markers are presented in 13. One should point out that position and informative value of ship identification signs and markers may vary greatly. Thus, identification signs may be located on board the ship and a sight line has to be positioned closer to the horizontal axis for inspecting such signs. At the same time, markers are located on the horizontal surface, and a sight line has to be positioned closer to the vertical axis for inspecting such markers. The required resolution of the SS determines the possibility of distinguishing between rather informative signs. This does not apply to color signs; however, color signs may lose their informative value under various weather conditions and lighting intensity. Image and videos processing algorithms, including those to solve search, detection and coordinate estimation tasks for objects of search of various types, are presented in detail in Therefore, a wide range of problems associated with an UAV landing on the ship are discussed in the known sources. However, some points must be further clarified. 2 Vol 9 (46) December Indian Journal of Science and Technology

3 Nikolay V. Kim, Nikolay E. Bodunkov, Ivan G. Krylov and Elena A. Fedyaeva 2. Problem Statement Scenarios, where communication between an UAV and the ship may be interrupted for a relatively long time and where the ship may change its path or velocity, for example, due to adverse weather are not considered in the above mentioned research. Let us consider implementation of the following stages under such conditions, according to tasks 1 and 2: guiding the UAV to a certain area where the searched ship is supposedly located and where ship path may change at random; searching for the searched ship. Let us assume that the secondary search for the ship is done, i.e. the communication with the ship was lost and ship coordinates may change essentially, while the UAV flies to the ship location. The question now arises of what conditions are required for the successful search for the ship or what conditions, when the UAV will not able to find the ship, are? This question may be considered as the question of physical realizability of ship search. It is obvious, that this search may not be realized, if the rate of search area expansion is higher than the rate of scanning of this area from the UAV. Let us take that and effective range of detection of the ship, 13 using the UAV SS for certain current conditions is, where q is a vector of current conditions. Let us assume that the area, where the searched ship is detected from the UAV is a circle with the UAV located in the center and this area equals to. (1) The search is complete, when the distance between the UAV and the ship is less than the effective range of detection.. (2) The secondary search area, where the ship may be located, changes subject to ship path and velocity, and - time of receipt by the UAV of latest ship coordinates or UAV flight time to the initial ship search area. Thus, for example, for the equally possible law of ship path distribution, maximum search area surface is determined by the area of circle:, (3) where time information missing (until search commencement), maximum velocity of the object of search (the ship). If the UAV started its search in, after ship coordinates were received, then the surface determines the area, where the ship may be currently located. Let us set theoretical UAV search performance 6,13 by:, (4) where UAV velocity. Then, the following will be the condition of physical realizability of the successful search:, (5) where current time, under the condition that the UAV is in the search area and the search has already started. If the condition (5) is satisfied, then the search time may be calculated by. (6) In order to satisfy the condition (5), search performance may be improved by increasing the flight velocity of the UAV. Accordingly, processing and analyzing performance of the incoming video information must be enhanced. Increased performance due to is limited by the conditions and functionality of the UAV SS. Reducing the time of ship search, which depends, to a certain degree, on the initial search area or time (3), is important. Thus, if the communication between the UAV and the ship is lost, UAV flight endurance to the search area should be minimized to reduce search efforts. 3. Results 3.1. Guidance methods Let us consider UAV flight to the ship search area. The effectiveness will be estimated according to the flight time. Guidance methods to determine UAV flight path that are similar to guidance of guided missiles, 16 will be further discussed in this research. To estimate effective application of these methods, UAV flights to the ship against transmission delays of different ship paths and different coordinates from the ship to an UAV were simulated. At the same time ship flight time was estimated and Vol 9 (46) December Indian Journal of Science and Technology 3

4 Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing ship search completion satisfied the condition (2). A guidance method means a given law of closure of a flying vehicle, in particular of a UAV to the object of chase. Subject to coordinates and motion parameters of the object, this method defines the required motion of the UAV, which ensures the UAV meeting the object. A theoretical UAV path, which is determined by equations of the guidance method, is usually called a kinematic or the required guidance path. One may distinguish between three fundamental guidance methods in the theory of guidance: Chase method, Fixed-lead guidance method, and Continuous-lead guidance method. The above methods essentially control the vector of UAV velocity subject to current position and velocity of the object of chase (the ship) Chase method In the chase method, UAV velocity vector turns for the whole flight time in a way to coincide with the vector that connects the UAV and the object of chase (the ship for our case), i.e. to reduce the angle between them to zero This method is easy to implement, as long as control is in proportion to the angle between the vectors. However, the UAV travels in this case to the point of current position of the UAV, which leads to increasing flight path and time. In the fixed-lead guidance method, UAV velocity vector must always be directed towards the target lead point. The lead point is calculated at the moment of ship detection and is considered fixed for the whole flight duration. At the same time, vehicle s velocity vector will maintain the constant angle (the lead angle) to the ship velocity vector, and the sight line will move parallel to itself for the whole flight duration. The lead angle is calculated by:, UAV path is straight-line almost along the entire length, which corresponds to the shortest route. However, this method implies an important assumption of ship movement to be straight-line and at constant velocity along the entire length. When this condition is not satisfied, the UAV will not come out into the ship area. As long as these assumptions are difficult to implement in real life, the continuous-lead guidance method is used. The sight line must constantly face the ship and the lead angle must constantly be re-calculated in this method. It is also assumed in the studied methods that the UAV gets uninterrupted information about ship location. However, in case of transmission delays major errors may occur leading to increased flight time. The following two methods will be compared in further studies: the chase method and continuous-lead guidance method Simulation and Experimental Data Let us compare methods of guidance of an UAV to the ship under different conditions. Time T of reaching the field of view of the ship is selected as a comparison criterion. Radius of field of view Simplified UAV models (including a material point) are used in many works devoted to this problem, for example 3,6. An overall view of an aircraft-type UAV that is used in the studies is presented in Figure 1. The UAV is described with a tertiary transfer function in the model. Limitations are imposed on UAV route change rate. Figure 1. A model of UAV. 4 Vol 9 (46) December Indian Journal of Science and Technology

5 Nikolay V. Kim, Nikolay E. Bodunkov, Ivan G. Krylov and Elena A. Fedyaeva The required route and velocity of the UAV are at the input of the model. Current coordinates, route and velocity are at the output. Velocity moduli of the ship and UAV are constant in all experiments and equal to and, accordingly. Ship motion along a certain path will be further simulated. The path is represented as a consecutive set of points. Ship coordinates are transmitted to the UAV. The UAV will make certain maneuvers to reach the field of view of the ship subject to the selected method and ship path. The chase methods and continuous-lead method for ship straight-line and uniform motion are compared in the first experiment. The ship starts its travel in the point A [30000, 30000] in the direction of the point B, i.e. along the horizontal axis. The UAV starts its travel in the point O [0, 0]. Ship coordinates are considered to be known to the UAV at any time. UAV flight simulations for different methods are presented in Figure 2. to the UAV at certain regular intervals. Delays may occur in transmission. Studying the effect of coordinates transmission delay in flight of the UAV is the goal of the next experiment. The ship changes its direction in the second experiment. It is assumed at the same time that ship coordinates are transmitted to the UAV following a certain delay. Let as consider a case of maneuvering (ship turn angle is at least 90 о ). The ship changes its direction every 300 sec. At the same time, the path is straight-line for the most part. Both methods involved 3 experiments each for UAV coordinates transmission delays dt of 0 sec, 180 sec and 600 sec (Figure 3). (a) Figure 2. UAV guidance by pure chase and continuous lead for straight-line ship motion. For the chase method, flight time is sec, and sec with the lead method. The flight time in chase is longer. This experiment is a simple illustration of our earlier conclusions. Thus, the lead method is most effective (in terms of flight duration) for the straight-line and uniform motion of the ship. This experiment is a simple illustration of our earlier conclusions. However, the ship may maneuver (change its motion direction). Ship current coordinates must be transmitted (b) Figure 3. UAV guidance with ship coordinates transmission delay. a) pure chase method, b) continuous lead method. An arrow indicated ship initial velocity direction in these figures. Turn maneuver of the ship at the beginning of travel leads to major errors in the lead method. Vol 9 (46) December Indian Journal of Science and Technology 5

6 Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing Simulation results are given in Table 1. Table 1. Simulation of UAV guidance with ship coordinates transmission delay dt, s T ch, s T pr, s Let us further discuss the case, when ship path changes every 300 sec (Figure 4). Table 2. Simulation of UAV guidance with the ship maneuvering dt, s T ch, s T pr, s For our next experiment, we will consider energetic maneuvering of the ship. At the initial time, the ship is in the point A, and its velocity vector is at 180 о angle to the x axis. Then, the ship turns and travels in the direction of the point B. In 1,200 sec the ship turns in the opposite direction. Simulated paths are shown in Figure 5. (a) (a) (b) Figure 4. UAV guidance with the ship maneuvering. a) pure chase method, b) continuous lead method. The ship travels from the point A, and vector direction is indicated with an arrow in Figures 4, a) and b). From these diagrams, UAV motion paths do not differ significantly for the delay of 0 sec and 180 sec. Let us discuss simulation results that are given in Table 2. (b) Figure 5. UAV guidance with ship coordinates transmission delay for intensive ship maneuvering. a) pure chase method b) continuous lead method. From Figures 5, the lead method produces a major error, because of ship turn at the initial time. 6 Vol 9 (46) December Indian Journal of Science and Technology

7 Nikolay V. Kim, Nikolay E. Bodunkov, Ivan G. Krylov and Elena A. Fedyaeva Total UAV flight time for different delays is given in Table 3. Table 3. Simulation of UAV guidance with ship coordinates transmission delay for intensive ship maneuvering dt, s T ch, s T pr, s Discussions Let us discuss the experimental data and make some conclusions, if guidance methods were effective (with view to the flight time). For straight-line and uniform ship motion, the lead method is most effective. The effectiveness of two methods for the maneuvering ship (see Figure 3) is similar. However, in case of coordinates transmission delay exceeding ship path change intervals, the effectiveness of the lead method drops abruptly. The lead guidance method proves less effective, in case of dramatic ship path change (turn by 180 о ). Also, with the delay increasing, the flight time for the lead guidance technique increases essentially, as compared to the pure chase method. Thus, one should point out that flight time sensitivity to the increasing delay of coordinates transmission from the ship to an UAV also appears lower for the chase method. 5. Conclusion Conditions of the successful implementation of ship detection in secondary search from an UAV have been determined; It has been shown that narrowing the initial area of secondary search depends greatly on decreasing UAV flight endurance against no information about current ship coordinates; Methods of UAV guidance to the search area of the ship, using the criterion of flight endurance for different ship paths and ship coordinates transmission delays have been compared; A greater sensitivity of UAV flight time to varying transmission delay duration, when using the guidance method that involves a deviated chase has been shown; and Conditions of effective implementation of the chase and deviated chase methods for different ship paths and transmission delay durations have been determined. 6. Acknowledgement The authors thank Ministry of Education and Science of the Russian Federation for supporting this research. The research was financed by the Ministry of Education and Science of the Russian Federation under the Grant Contract dated 27 October, 2015, (Grant Number , Contract Unique Identifier RFMEFI60715X0127), the grant was assigned for the research into Development and implementation of autonomous light-weight aircraft UAV shipdeck landing using machine vision system. 7. References 1. Kim NV, Hyun YM, Yang HK. Performance analysis of aircraft automatic landing system based on surface image processing. Proceedings of The World Congress of Korean and Korean-ethnic Scientists and Engineers in Seoul, Korea Frølich M. Automatic Ship Landing System for Fixed-Wing UAV. [Master s thesis]. Trondheim : Department of Engineering Cybernetics,Norwegian University of Science and Technology Kai W, Chunzhen S, Yi J. Research on Adaptive Guidance Technology of UAV Ship Landing System Based on Net Recovery. APISAT2014, Asia-Pacific International Symposium on Aerospace Technology. 2014; 99: Carrier Phase Compensates for Wind and Wave Motion. UAV Shipboard Landing with RTK. GPS World staff Available from: 5. Chaves SM, Wolcott RW, Eustice RM. NEEC Research: Toward GPS-denied Landing of Unmanned Aerial Vehicles on Ships at Sea. Naval Engineers Journal. 2015; 127 (1): Kim DP. Methods of Chase and Search for Mobile Objects. Moscow: Nauka Nayak G C, D Souza V. Optimized UAV flight mission planning using STK and A* algorithm. International Journal of Emerging Technologies and Engineering. 2014; 1(5): Gao X-Zh, Hou Z-X, Zhu X-F, Zhang J-T, Chen X-Q. The shortest path planning for manoeuvres of UAV. Acta Polytechnica Hungarica. 2013; 10(1): Vol 9 (46) December Indian Journal of Science and Technology 7

8 Selecting a Flight Path of an UAV to the Ship in Preparation of Deck Landing 9. Clark S, Goodrich MA. A hierarchical flight planner for sensor-driven UAV missions IEEE RO-MAN: The 22nd IEEE International Symposium on Robot and Human Interactive Communication,Gyeongju, Korea. IEEE; 2013 Aug p Advanced Path Planning and Collision Avoidance Algorithms for UAVs[PhD thesis] july 26. Available from: Date accessed: 26/07/ Forsmo EJ. Optimal Path Planning for Unmanned Aerial Systems.[ Master s thesis]. Norway:Norwegian University of Science and Technology. Department of Engineering Cybernetics; Abramov N S, Makarov D A, Khachumov M V. Controlling Flight Vehicle Spatial Motion along a Given Route. Automation and Remote Control. 2015; 76(6): Kim N, Bodunkov N. Adaptive Surveillance Algorithms Based on the Situation Analysis. In: Favorskaya M, Lakhmi C J. (Eds.) Computer Vision in Advanced Control Systems: Innovations in Practice. Springer. 2014; 2: Forssyth DA, Ponce J. Computer Vision: A Modern Approach. 2nd ed.. New York: Pearson Education Inc; Jähne B. Digital Image Processing. Springer Yanushevsky R. Modern Missile Guidance. CRC Press, Vol 9 (46) December Indian Journal of Science and Technology