a survey on marine control systems

Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 A Survey on Marine Control Systems Tổng quan về hệ thống điều khiển hàng hải Hung Duc Nguyen University of Tasmania / Australian Maritime College e-Mail: nguyenhd@amc.edu.au Abstract In this paper, a survey is made on modelling, simulation, control design, advances, achievements and trends in marine control systems. An overview of history of development of marine control systems is outlined. Over a long history, many achievements on marine control systems have been reached in both theory and practice. With the aid of computers and high performance software many complicated control algorithms could be applied in modelling, simulation and design of control systems for marine vehicles including surface vessels and underwater vehicles. The development of GNSSs (GPS, GLONASS and GALILEO) and RTK/D-GNSSs stimulates design of accurate, precise and high-performance control systems for marine vehicles. Telecommunication satellite-based broadband techniques are a trend of remote control systems at seas. The paper discusses challenging problems in design and simulation of marine control systems. The paper also deals with some potential research projects related to the marine control engineering at AMC/UTAS. Tóm tắt: Trong bài báo này tác giả trình bày tổng quan về mô hình hóa, mô phỏng, thiết kế điều khiển, những tiến bộ và thành tựu cùng các khuynh hướng phát triển hệ thống điều khiển phương tiện trên biển. Bài báo khái quát lịch sử phát triển hệ thống điều khiển phương tiện trên biển. Qua lịch sử lâu dài cho đến nay có nhiều thanh tựu trong hệ thống điều khiển hàng hải. Bằng sự hỗ trợ của máy tính và phần mềm tính năng cao người ta có thể áp dụng nhiều thuật toán điều khiển phức tạp trong mô hình hóa, mô phỏng và thiết kế hệ thống điều khiển cho phương tiện trên biển. Sự phát triển của các hệ thống vệ tinh dẫn đường toàn cầu (GPS, GLONASS, GALILEO) và hệ thống định vị vệ tinh vi phân đã kích thích việc thiết kế các hệ thống điều khiển chuẩn xác, chính xác và có đặc tính tốt cho phương tiện trên biển. Các kỹ thuật dải băng thông rộng thông qua vệ tinh viễn thông là một trong những khuynh hướng phát triển hệ thống điều khiển từ xa trên biển. Bài báo thảo luận về những vấn đề thách thức trong thiết kế và mô phỏng hệ thống điều khiển hàng hải. Bài báo cũng đề cập đến một số đề tài nghiên cứu khả thi liên quan đến lĩnh vực công nghiệ điều khiển hàng hải tại AMC/UTAS. Nomenclature Symbol Unit Meaning ν   T u,v,w,p,q,rν η   T n,e,d, , ,   η Abbreviation AMC Australian Maritime College UTAS University of Tasmania PID Proportional, Integral, Derivative LQG Linear quadratic Gaussian GPS Global Positioning System GNSS Global Navigation Satellite Systems DP Dynamic positioning D-GPS Differential GPS RTK-GPS Real-time Kinematic-GPS IFAC International Federation of Automatic Control ECEF Earth-centred Earth-fixed frame ECI Earth-centred inertial frame NED North-East-Down frame FPP Fixed pitch propeller CPP Controllable pitch propeller 1. Introduction Marine control engineering is about applications of control theories into marine and offshore systems. It involves the research and development of new control algorithms, hardware and software for control systems in maritime engineering systems. Marine transport is more cost-effective than other transports. The world’s fleets carry the majority of cargo. In many countries like EU, Australia, America, Japan and Korea the number of seafarers is decreasing because sailing at sea is a job in severe working conditions. This requires a high-level automation on board cargo carrying marine vehicles because the shipboard high-level automation can reduce the number of crew. Advances in computer and information technology, data communication technique and instrumentation engineering play a very important role in development of new control solutions for optimal and high-performance control systems and fuel saving. The new control solutions are based on modification of feedback control algorithm and new configuration of hardware. The building of new types of marine vehicle and craft inspires new design of instrumentation and control systems. In recent decades, more and more ROVs/AUVs have been applied in exploration of seabed, discovery and 115 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 exploitation of marine resources. This requires new solutions for data communication and control algorithms. Control of ROVs/AUVs is a great challenge because they are operating in 6-DOF. This paper is organized as follows: Section 1 Introduction; Section 2 Current status of marine control systems; Section 3 Kinematics and kinetics; Section 4 Overview of marine control systems; Section 5 Modelling and identification of marine vehicles; Section 6 Experimental facilities; Section 7 Challenges, Section 8 Trend; Section 9 Potential projects at AMC/UTAS; and Section 10 Conclusions. 2. Current Status 2.1 Overview of History The invention of the gyroscope contributed much to the development of a ship’s autopilot system. The development of the electronically-driven gyroscope was motivated by the need for more reliable navigation systems in steel ships and underwater warfare [3][4]. The successful design of the gyroscope at the beginning of 20 th century was the key breakthrough in automatic ship control since it led to the development of autopilots and other control systems (see Fig. 1). Fig. 1 Diagram of history of marine control systems 2.2 Research Activities The IFAC organizes every 3 year (triennial) conferences on marine systems including CAMS (Control Applications in Marine Systems), MCMC (Manoeuvring and Control of Marine Crafts). The scopes of these IFAC conferences on marine control systems are broad ranges from autopilot to dynamic positioning systems and various applications of control theories in control, simulation and modelling of marine vehicles. These IFAC conferences on control of marine vehicles cover a wide range of scopes, for example, ship manoeuvring, autopilots, roll damping, dynamic positioning, automatic mooring and anchoring, navigation, guidance and control of autonomous surface and underwater vehicles, operational safety etc. 2.3 Development of GPS/GNSS and IMU/INS Since 1995 when the GPS became operational for civil use, the accuracy of GPS/GNSS has been improved significantly. The augmentation, integration and availability of GPS, GLONASS and GALILEO for civil use with high accuracy, precision and reliability inspire engineers and researchers to design new types of tracking and path-following control system. Moreover, the development of IMU/INS and integration of GNSS and IMU/INS allows more accurate and precise navigation systems to be designed and helps more complicated marine control systems to be developed. 3. Kinematics and Kinetics of Marine Vehicles 3.1 Reference Frames In the design of marine control systems, some reference frames for descriptions of kinematics and kinetics of marine vehicles are often used. Fig. 2 shows Earth-centred reference frames (the Earth- centred Ear-fixed frame x e y e z e , and the Earth-centred inertial frame x i y i z i ), and geographic reference frames (the North-East-Down coordinate system x n y n z n and the body-fixed reference frame x b y b z b ) [3][4]. Fig. 2 The ECEF frame x e y e z e is rotating with angular rate with respect to an ECI frame x i y i z i fixed in the space [3][4] Fig. 3 shows the 6DOF velocities in the body-fixed frame. Table 1 gives the notation for the 6DOF motions, forces and moments, linear and angular velocities, position and Euler angles for marine vehicles. z i , z e ω e y n x n z n BODY y x z NED ECEF ω e t y e x e y i x i 116 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 Fig. 3 The 6DOF velocities u, v, w, p, q and r in the body- fixed reference frame x b y b z b [3][4] Table 1 The notation of SNAME (1950) for marine vessels 3.2 Equations of Kinematics Referring to Fig. 2 the 6-DOF kinematic equations in the NED (north-east-down) reference frame in the vector form are,    η J η ν (1) where       n b 3 3 33        R Θ0 J η 0TΘ (2) with 33 Sη and 3 ν . The angle rotation matrix   n 3 3 b  R Θ is defined in terms of the principal rotations, x, 1 0 0 0 c s 0 s c            R , y, c 0 s 0 1 0 s 0 c            R and z, c s 0 s c 0 0 0 1             R (3) where s  = sin(.), c  = cos(.) using the zyx-convention,   n b z, y, x, :    R Θ R R R (4) or   n b c c s c c s s s s c c s s c c c s s s c s s s c s c s c c                                         R Θ (5) The inverse transformation satisfies,     1 n b T T T b n x, y, z,     R Θ R Θ R R R (6) The Euler angle attitude transformation matrix is:   1 s t c t 0 c s 0 s /c c / c                   T Θ    1 1 0 s 0 c c s 0 s c c                 T Θ o 90   (7) It should be noted that    T Θ is undefined for a pitch angle of o 90   and that     1T  T Θ T Θ . 3.3 Equations of Kinetics Referring to Fig. 3 the 6-DOF kinetic equations in the body-fixed reference frame in the vector form are,       0 wind wave       Mν C ν ν D ν ν g η g τ τ τ (8) where M = M RB +M A : system inertia matrix (including added mass);   C ν =     RB A C ν C ν : Coriolis-centripetal matrix (including added mass);   D ν : damping matrix;   g η : vector of gravitational/buoyancy forces and moments; 0 g : vector used for pretrimming (ballast control); τ : vector of control inputs; wind τ : vector of wind-induced forces and moments; and wave τ : vector of wave-induced forces and moments. 3.4 Equations for Manoeuvring of Surface Vessels For surface vessels their motions are often limited to 4-DOF: surge, sway, yaw and roll. It is assumed that the vessel is symmetric about the plane of XGZ and the origin and the mass concentration at the centre of gravity, four 4-DOF kinetic equations are expressed as [13], mv mur Y (9) mu mvr X (10) zz I r N (11) xx IK (12) where m is the mass of the vessel; I zz is the moment of inertia about z-axis; and I xx is the moment of inertia about x-axis. X, Y, N and K are forces and moments acting on the vessel, including propeller-generated forces and moments, hydrodynamic forces and moments due to interaction between the propeller and the hull, rudder- 117 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 or control surface-induced forces and moments and external disturbances. Equation (1) is simplified as, pos x ucos vsin    (12) pos y usin vcos    (13) 3.5 Equations for Environmental Disturbances Environmental disturbances include wind, waves and currents. According to Fossen [3] for control system design it is common to assume the principle of superposition when considering wind and wave disturbances. With effects of external disturbances Equation (8) is rewritten as,         RB RB A r A r r r r 0         M ν C ν ν M ν C ν ν D ν ν g η g τ w (13) where wind wave w τ τ and rc ν ν ν (where 6 c ν is the velocity of the ocean current expressed in the NED). Further information on modeling environmental disturbances can be found in [2][3][4]. 3.6 Discrete-time Models for Marine Vehicles The classical methods of designing control systems are using continuous-time models including differential equations, transfer functions and state- space models. The computer-aided methods are using discrete-time models, including difference equations, pulse transfer functions and discrete-time state space models. Auto-regressive models are often used for stochastic control algorithms and model reference control. Discretisation of the following continuous- time state-space model (14) results in (15) or (16) where (17) (18) For stochastic control systems the following auto- regressive average moving exogenous model and auto-regressive exogenous model are used: (19) (20) 4. Overview of Marine Control Systems – Motion Control Motion control of marine vehicles involves the guidance, navigation and control of:  surface vessels;  underwater vehicles including submersibles and submarines; and  oil rigs, floating and subsea structures. The motion control systems for marine vehicles include ship autopilots, roll damping/stabilising systems and dynamic positioning systems. For surface vessels the desired motions are surge, sway and yaw (turning) while undesired motions are heave, roll and heel, pitch and trim. Surge, sway and yaw motions are often controlled by a rudder or control surface, FPP or CPP, side thrusters. The undesired motions are reduced to an acceptable level by some motion control strategies such as fins, trimtabs, interceptors, T-foils, rudder-roll, lifting foil and air cushion support. 4.1 Guidance, Navigation and Control of Marine Vehicles An entire modern control system for marine vehicles has three subsystems as shown in Fig. 4 [3]:  guidance system;  sensor and navigation system; and  control system. Fig. 4 The GNC signal flow [3] The guidance system is used to generate desired signals based on the prior information, predefined trajectory and weather data from weather forecast stations. Some techniques that are applied in the guidance systems are target tracking, trajectory tracking, path following for straight-line paths, and path following for curved paths [3]. The sensor and navigation system consists of necessary sensor and navigation devices such as GPS/GNSS receivers, wind gauges, depth sounder, speed log, IMU/INS and engine sensors. In order to have “clean” data for control purposes observer, filter and estimator techniques are applied. The control system is where control algorithms are synthesised and control signals are computed. Modern control algorithms are applied. Fig. 5 shows an example of recursive optimal trajectory control system. Fig. 5 The GNC signal flow of the recursive optimal trajectory tracking control system [7]             k 1 h kh (k 1) exp h k exp k 1 h k d         x A x A Bu x Ax Bu       k 1 k k  x Φx Δu   exp hΦA   1 Δ A Φ I B             1 1 1 z k z k z k    A y B u C e           11 z k z k k  A y B u e 118 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 As shown in Fig. 5 the control system consists of a guidance system that generates desired course, speed and course changing points based on the LOS, waypoint and decay exponential techniques. The sensor and navigation consists of GPS/IMU/INS, gyrocompass, sensors and a recursive estimator. The control system consists of a controller based on the optimal control law. 4.2 Autopilots Autopilots are used for course keeping and changing. The common method for conventional vessels equipped with a propeller and rudder is illustrated in Fig. 6. As shown in Fig. 6 the course (yaw) angle and yaw rate are measured by a compass and gyro. For a waterjet-propelled vessel, the course is controlled by the waterjet nozzle. Fig. 6 Ship’s autopilot system [4] Modern and intelligent control algorithms have been applied in the autopilots. Fig. 7 shows an example of a stochastic model based autopilot with a combination of a recursive estimation algorithm and the self-tuning control algorithm. Fig. 8 shows an example of the neural networks-based autopilot. Fig. 7 Ship’s recursive self-tuning autopilot system Fig. 8 Ship’s neural networks-based autopilot system 4.3 Rudder-roll Stabilisation Systems The roll motion of a marine vehicle has bad and unexpected effects on crew and passenger heath and cargo as well as the stability of the vehicle. The effects of roll motion (especially the parametric roll motion) are seasickness, damage of cargo and damage of vessel. A rudder-roll reduction system is based on the principle illustrated in Fig. 9 and Fig. 10. The main requirements for this system are:  fast rudder slew rate;  accurate measurement of roll motion; and  low pass filters. Fig. 9 Principle of a rudder-roll stabilisation system Fig. 10 Autopilot system with rudder-roll reduction Fig. 11 shows an example of responses of an autopilot system with rudder-roll damping function. Fig. 11 Responses of an autopilot system with rudder-roll reduction 4.4 Dynamic Positioning Systems Dynamic positioning systems are used to control marine vehicles at very low speeds where the effect of rudder or control surface is almost zero. A modern DPS has many functions such as autopilot, dynamic positioning, trajectory tracking and shifting anchor alarm. To design a DPS the waypoint, LOS and decay 119 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 exponential techniques are applied. Fig. 12 shows the main forces and moments generated by actuators and external disturbances on a vessel equipped with a DPS. Fig. 12 The main forces and moments for DPS design (courtesy of Kongsberg) In the DPSs there are more than two controls. DPSs require network data communication buses. Modern and intelligent control algorithms such as optimal control, self-tuning control and fuzzy logic control have been applied in the design of DPSs. 4.5 Networked Control Systems and Integrated Bridge Nowadays marine control systems are in forms of a networked control system, distributed control system and integrated bridge that allow the operator to control many onboard systems. The networked control systems have data communication buses such as NEMA, CANOpen, and Profibus. Fig. 13 shows a networked control system with NAMA data communication devices. Fig. 13 Concept of networked control system with data communication bus (NEMA) The centralised control systems are obsolete and replaced with distributed and networked control systems. For high-level automation marine vehicles a networked control system has some main features: integrated, distributed, supervisory, redundancy and safety as shown in Fig. 14. Fig. 14 Example of high-level automation control system on a modern vessel (courtesy of Kongsberg) 4.6 Control Systems for ROVs/AUVs, Oil Rigs and Floating Structures Control of ROVs/AUVs, oil rigs and floating structures is a greater challenge in comparison with control of surface vehicles because of their complexity, moving at low speeds and underactuation. Control algorithms and methods for ROVs/AUVs are described in [3][4][11] and [12]. 5. Manoeuvrability, Modeling and System Identification of Marine Vehicles (Hydrodynamics) To assess manoeuvrability of marine vehicles is important for safe operation. The manoeuvrability of ocean vehicles must meet IMO standards, including interim standards for ship manoeuvrability IMO Resolution A.751(18), 1993 and standards for ship manoeuvrability IMO Resolution MSC137(76), 2002, issued by the IMO Maritime Safety Committee. The marine vehicles built with very poor manoeuvring qualities will cause marine casualties and pollution. The manoeuvrability is often related to the:  seakeeping: a measure of how well-suited a marine vehicle is to conditions when underway; and  seaworthiness: the ability of a marine vehicle to operate effectively under severe sea conditions, i.e. very good seakeeping ability. To quantify the manoeuvrability is to identify hydrodynamic coefficients of the manoeuvring models. Its applications are:  manoeuvring characteristics (for various manoeuvres);  stability assessment; 120 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011  computer and HIL simulation (full mission manoeuvring simulators) for educational and training purposes;  control design (stochastic control, model based adaptive control);  fault detection and diagnostics; and  prediction of forces and moments due to the interaction between many submersible bodies. The quantitative representation of manoeuvring characteristics of marine vehicles consists of straight- line stability and directional stability. The methods to assess the manoeuvring characteristics are the turning circle test, Kempf’s zig-zag test, Dieudome’s pull-out manoeuvre test, Bech’s reverse spiral manoeuve test and stopping trial. Many authors proposed manoeuvring mathematical models, for examples, Abkowitz (USA: SNAME), MMG model group in Japan (SNAJ, JTTC), Norrbin (1970), Blanke (1981), Nomoto and Sons, etc. Further information can be found in [3][4][5]. The most common and well-known model of manoeuvring is the Nomotor’s first order model that relates the rudder angle and yaw rate (turning rate): Tr r K   (21) where T and K are manoeuvrability indices. In order to quantify the manoeuvring characteristics of marine vehicles and determine hydrodynamic coefficients of the manoeuvring mathematical models, it is necessary to conduct full-scaled or model-scaled experiments as shown in Fig. 15. Fig. 15 Experiments for prediction of hydrodynamic coefficients In order to estimate hydrodynamic coefficients of a vehicle there are several methods among which the following are widely used:  Recursive least squares algorithm; and  Recursive prediction error method 5.1 Recursive Least Squares Algorithm (RLSA) The recursive least squares algorithm is based on the least squares algorithm proposed by Gauss. This method is illustrated by the flowchart in Fig. 16. Fig. 16 Flowchart of RLSA 5.2 Recursive Prediction Error Method (RPEM) The recursive prediction error algorithm was proposed by Ljung based on the Kalman filter and is illustrated by flowchart in Fig. 17. Fig.17 Flowchart of RPEM 4.7 Fault Detection and Diagnosis Monitoring and Supervision and Fault Tolerant Control Recursive system identification methods are applied in fault detection and diagnostic monitoring and supervision of marine and offshore engineering systems. They are also applied in fault-tolerant control. The conceptual system of fault detection and diagnostic monitoring and supervision is shown in Fig. 18. The fault detection system requires prior knowledge of the plant (theoretical data) and sensors to collect actual data. The system compares actual data with the theoretical data and thus detects any 121 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 faults occurring in every component of the engineering systems when there is a great difference between two sets of data. The system provides solutions to manage faults. Further information on fault detection and diagnostic monitoring and supervision can be found in [14] [15]. Fig.18 Concept of fault detection and diagnostic monitoring and supervision for marine and offshore systems 6. Experimental Facilities In order to support control design and to realise marine control systems it is necessary to utilise experimental facilities for full-scaled and model- scaled experiments. Experiments require the following facilities:  physical models or prototypes of marine vehicles;  model test basin with artificial wavemaker and wind generators for free-running models;  towing tank with PMM for captive models;  full-scale vessels (expensive); and  control hardware (instrumentation electronics, data communication) and software. The AMC/UTAS possesses the world’s leading maritime experimental facilities. The facilities include the towing tank (see Fig. 19 and Fig. 20), model test basin (see Fig. 21) cavitation tunnel (see Fig. 21), and circulating water channel (see Fig. 22), full mission ship manoeuvring simulator, dynamic positioning simulator, and training vessel (Bluefin). Fig.19 AMC Towing Tank Fig.20 AMC Towing Tank with PMM and captive model Fig.21 AMC Model Test Basin with wavemakers and models Fig.21 Three dimensional view of the AMC Capvitation Tunnel Fig. 22 The CWC and its arrangement Other institutes that also have the world’s leading maritime experimental facilities are Norwegian University of Science and Technology and MARINTEK, Tokyo University of Marine Science and Technology. 7. Challenging Problems In design and simulation of marine control systems some challenging problems are:  underwater communication between the AUVs and mother vessel;  energy for ROVs/ AUVs that operate underwater for a long time;  fault detection and diagnostics and safety, this leads to losses of expensive ROVs/AUVs  control and operation of ROVs/AUVs at very deep waters;  watertight electronic components; and  in-door navigation techniques for experiments. 122 Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011 VCCA-2011 8. Future Trend Recent trends show the following applications:  networked control systems with data communication buses;  Internet-based control systems utilising satellite broadband services;  applications of advanced and intelligent control algorithms;  wireless network;  underwater acoustic navigation systems for ROVs/AUVs; and  optical communication between ROVs/AUVs and the carriage vessels. Fig. 23 shows an example of remote control system via satellite broadband services in Norwary. Fig. 24 shows another example of remote control system via satellite broadband services in Japan. Fig 23 Remote control system via satellite broadband services (Norway) Fig 23 Remote control system via satellite broadband services at Tokyo University of Marine Science and Technology, Japan 9. Potential Projects Related to Marine Control Engineering at AMC/UTAS The AMC, possessing the world’s leading maritime experimental facilities, is undergoing several potential projects related to marine control engineering. These projects are:  design and testing of ROV/AUVs;  modelling, simulation and control of ROVs/AUVs;  modelling, simulation and control of AUVs using a cyclic and collective pitch propeller;  modelling and control of surface vessels with electrically-operated water-jet (GreenLiner)  development of ROVs/AUVs with a collective and cyclic pitch propeller;  development of a (solar-wind-diesel) trybrid trimaran and its control systems;  development of automatic manoeuvring systems for surface vessels;  development of dynamic positioning systems by applying advanced control algorithms; and  prediction, simulation of hydrodynamic interaction between many submersible bodies. 10. Conclusions The paper has discussed the current status of marine control systems and description of kinematics and kinetics of marine vehicles for design and analysis of their control systems. It has overviewed marine control systems and modelling and identification of marine vehicles. To design and analyse control systems full-scaled and model-scaled experiments are necessary and require maritime engineering specialised experimental facilities such model test basin, towing tank, circulating water channel. The paper has also dealt with future trend of marine control application and some potential projects at AMC/UTAS. References [1] Roberts, G.N. and Sutton, R (Editors). Advances in Unmanned Marine Vehicles. The Institute of Electrical Engineers, 2006. [2] Fossen, T.I Nonlinear Modelling and Control of Underwater Vehicles, PhD Thesis. Norwegian Institute of Technology, 1991. [3] Fossen, T.I Handbook of Marine Craft Hydrodynamics and Motion Control. John Wiley and Sons Inc. 2011. [4] Fossen, T.I Marine Control Systems – Guidance, Navigation and Control of Ships, Rigs and Underwater Vehicles. Marine Cybernetics, Trondheim, Norway, 2002. [5] Fossen, T.I Guidance and Control of Ocean Vehicles. John Wiley and Sons, 1994. [6] Wadoo, S.A. and Kachoroo, P Autonomous Underwater Vehicles: Modeling, Control Design, and Simulation. CRC Press, 2011. [7] Nguyen, H.D Multitask Manoeuvring Systems Using Recursive Optimal Control Algorithms. Proceedings of HUT-ICCE 2008, pp. 54-59 Hoi An, Vietnam, 2008. [8] Nguyen, H.D Recursive Identification of Ship Manoeuvring Dynamics and Hydrodynamics. Proceedings of EMAC 2007 (ANZIAM), pp. 681-697, 2008. [9] Nguyen, H.D Recursive Optimal Manoeuvring Systems for Maritime Search and Rescue Mission, Proceedings of OCEANS'04 123 . h46" alt="" src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAKsAAADkCAIAAABohEP3AAAACXBIWXMAABYlAAAWJQFJUiTwAAAgAElEQVR42sy9y44kS5IlJkdE1czc45WZ91HVXT3d0zPEAMRsCK4IcElyy0/hL3DBT+KSAP+AOy64I2Y4013ddR+Z8XB3M1ORw4WamZs/IjMy69bMOBKByAgPd3NTUXkcOXIU/8v/+r/97//H/ykioEwPD6EoICIARIRcfieEiEhAAAAI4fKE5RWUIkERgYjVJ1LcnRCaAnCVEJI0gUGVYiEhdMj6Ud+ivn59i3pVl4/62+VJFKnPq5cUEVi9ICHrN1IeP2B9PkWgqJ9xuQkRcfamFleuYf04u1ouV4j5vxCl/PmPy2tT1fV9C4Rgvimrx//8P/2PSf6MKyApuPLJ10u4vD 4A6 rKoBHBuWJTPrO6f81DVtQWEUL70sUkKQE7XiQvL+01W7rd64JWNsb5ekbj6i8TZLo9L9Tajm+7m/LeTz+Dxz9cbiBBRKMD5h7xYA76+9idb/Ks//PlLEa9+SGL1fK8Lz/o2Kv/lPi5vwtGrVQsW4SufOf2G17FsC1BkXumYr08JYv55XQWerAQhn/fwf46vwuwSAVBE32Tnv/lG/M9mHxAJ4bVNJkmWKPg1PmCJ0LOTfPV2LN+4EIJ5OWNZ7 +oN4 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 3a+ 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 5A1 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 +a2 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 7A/ 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 /a+ Eq2oCI+qLgZPUIg2qEhK86/If3t3ft83395vOKOGqPm4bbvLf/Pjhx3e3XUKTMFKapklqCAz9AMA05aYRtVnth6vgRpKH/b5uU4r4rPMZHkEueUDEhFYdSe6QoEMI1YhQUVOEaEAjxCNyziGuWgvCFiIxFLQGSmFQJOckI2tavn9+aZrOUmo3N0g2FicIWHWRZ/zxM8TFGTFPiGJRxQRk4vhwkX25dHjp7fv7M5JPbxFLqJtx+TNf5RY2rTRqB3mebhFAVQJkgjRJH7abu 7a9 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 +a/ /WalBZH+Y6l4hfEQWuVIaKxcMsR3ZfGf 9a0 0FmsTKpDILVmObyhGl2yVmTXy+kCVdTr4CoKETAyGqbnDvTRDawZDVvYFaAHkGYNW0HlpSypZRSo6KmpppMp1NoIijuMuu7r8Sujy6hRFDGNtX4TXdnRJ1nn+dHahqvdEwqtwiq1Akzp4R4xOiUru1KKQq5udncH242XWO7nUpIqNhKqCzmKR5TiLl71GMz9Jx/y1NtHVy0f 9a0 5fUZ7GsyL 1a6 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 0A/ DODhFJB32++JlZjN48ZEeFYhzp8MFVm+6KW 7a9 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 /a4 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 8a3 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 8A+ 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 9A3 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 /A0 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 1a5 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 7a4 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 9a3 9ZLhxeHZZWAAnavGHXGfHMgS0hGbgRBje+pMMePDiDDnd+o8i1QZOjoj2qYOYCb5kOi1m1WihKkGNEoIaYeDMwojsnilmECGYIAgmQyAnBhEJI+VqcZgPOXs1ZUoISYEk427r0yjDNIzDIOApZZaUOV+sV2fTKr76zryfzeqh1jxNq9VYzbf7fVErakTE6xEiVQ0sxFJMSw0PDbJCSgNJoqrVyayNRpyTZEgWCJFbC0riIDYEgyiCQ9JB59mFmMMbGRK8pIQsmq1FENBHMb3VvkBo4zgbPnXtLGl3p2kO7UbRLijt/2LvKLwvvR9wPAW4z1WWBdtyNF93Vjl5+AKcaxHzTdLfgnuQWBocnYEM4SDm1Iae5l6rzqUQQoPY3VWTMySEJI1wHpxTBME0tM62Jy2D4GzaDGDwVM0DaEDXs/V6rlVKNTN1c/MkQhRqGlrDLNzmWg+lSgApk2i4c3Db7QCIUGsRBsfSJAwRqUZG5AGLqPZGau07GsAvsXPzZXaE4vbvfDvECQ9AsW61t9qa5eAW0IN+F0B4tOSERAzGSGThEi7MwoIwSalNBJSiujljq+ZGQaRuLAakcRiu5uv09Aru5O568DrDY8rDo83ZZlx98N77ksTU9vOh0xskDavVOE0i0sQEAVS1MJ0P81xmc396dVkOh7PVKCJpGFqgKYuIjJwT9/RXBvtyrDJz0lJVpEYcVGtYvP6Zuqkb3lwEfGuh3KR537Uafnkr 4A6 BCicFI9PLIFTdfXyTZSaB1JyFy0lvbddnNDBYBJF6hEUS6WYSFnc3CnfzCJe8r6YgNava3jJ123nVUIUZgzhqRk2cs/DvPnkqTj/72S9zYveoRcGhGpiGPI2r9TpPqydPnjx6/JhYDqWUUi6vt9fX18V0b6WU2RGTSJv2ZkkpDZIGIlRDaHiTSPd+W9/YHDy77Wo1 4A1 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 8A1 NwsiD24iU7IzNTDEcVtNi9EunA7v6wn+3N 1A/ op8HrledxIgPqN8KHv1+nd7n16sTAO414M4it+ORrM31vgSAJrOMwr1LIMkjg44OTkS2uhKcdrqUw+gC+mKQ/+iGUY1jmNZdY6l3qumfhsHM5WYwLzsDaZ2o56KOWzzz57+vSzWursOuZRcuaIaVoTE4QB0rCitWqBcMp5mDKYyXW3/VQoBuYcSAB35SRRsJoeZlUNTinLwEmCqKodSjEM+1quDodDuDa2DsWrXX+3sx+/zCluRCS/Vwfc+Z9MS5/yDqL03kv0nhN3c+07Yg/eZL7hoGAxDyVSOom /a9 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 0a4 MDXIeEe5q1gMYomWQpdTyqxOnBKR1kKchSUosKQhaneC1ViJIztm9NheIG9r73QzypuruJMRCKWsNVb/cH5w4JS5mLIh7fZKbq3Kn+7PH6W1wYcX2wDgs/vaX7gPxGn7Ay3cPvMHpcL9l8YV7170q5haN29UESpECTvAgC6/91hEaZIiT4BUgqOeGV6dGmolIzMyRzterKXLeM233ez0UPWgrE6yFzQqngIdFeARq0uDYrabx/GIzpGE1DMMg45BTm0KFqoV7M4BZEEkaAl69Cdu9z+3cFcHCrlTUQlKIXB8Ol1fXIZKG8VCVM99Pleki9yXPLyJwt/viWKQG7Sy3Vx/keN0KuHniF 1A4 gv7/vp/e7i54ABTWZWVwciOyoBLeRHRHNZWIsHc/pbcT15qjG+EuTAwmcBYkXjPROPD1zq6reXGts7m1YN+QvuCDwSxnq3yxmR49ebRZjavVOE45Nyl/mBuFGkUY0dwD6VhEgtkW7VXrCEGaHiRKqRQc43B1eX1QC0nHe 5a9 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 +a5 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 0a4 HB4HIkT3t0OnIt1WrszgJTKkIJgLAxOdJghyKtV0d32sC1V3aOEBTEEzALifv3qx3kHMQVLqapODhRT5HEZ2/QET8drPHoPJsbFHW0+Hp 7A0 uI/vvmMcK/VA0p0qvt4eYzZF7Ek3T7mv4hubEERHGl8Zh6Nf+XugWOsRUP3gDuhoEXxWlCppghj5gj2aMOA3mlplF8Is4jPgBAhnIg4mWV1SU29HqEUCK 5a1 OZ6cCKWFAEGCUviFo2H0 8A1 Ym7avqpatYqIoJGC7VaUJ8krCqb7E+EFKPdmb93tSc3L6sn06nCaf0be/U1P+5/zLXoYQf9Ug5pOrLlEXC0rz7miAAAgAElEQVSCcXw3wUSBtsuGaiBxbcpViwWPwM21BEJwb8kxEbkLMzebF1EhH8LZIAsR28K91CCoO5hTkHkwuQhzOJlSJHBjSouBWCTA+7lqQI3UHcNQXTmkvYZXHPwPwsButMLMb7qvvtkfSyC6HUryADmg5d7EQ/VbnGgJX70HfIFl0P9DbyyiAMXq7ELqQU2FSJBA1oQFDe67TFFzrdUjyCKYIkIa8VlaMFpw9B67u7cTGYjESKmDl5eQ+WB3iuZf8cY9UXWW9jTTWjgzSxCDEpOwsGRmJneRwQhF68WTdx7tDr/5zW+NOSQ3sHKjxHUzRbxmBRzjpW96J28WIXI/4uFVs8GbDeeh/k/bVO 8A4 2/tMEcBMm5dbU/+SngT7Trdbn80h7m5hVtzkm/OzuuVH6oCIERmdrdjavpp/LGdjF3DI+BMHGCGKC1qAEYEh3ufJVM08kBQuJNwkKHdbgBEqEe4R2YCI3MaMydIYhGWxIkg4aFmnBIzz9Wq4Z13v/ZO1b/75NN23pAfr7zh7W2jk9j2uPWpn0YG9z497mr1XraX3Pn6q3KH43XbhYMkblz/d/eAe4Ey/RWHv8lnf79dePqi3ftsou1+DH73 3a9 dpqRuQpHgroVvCpveDgkiESZKxx3II0jdBYze1Ft+HBNAIuYu3b/cJYfFNZEs47gOjGZwSnlMOXMSpiSRwAwWERCIkfPoxLNWiiCW9cXjx195b315mfKozNVs4GGhDUYcc7tvqz8e/LSOG8BpoPCNdjTuPvf302VfvgJuz3O/GK4OC/J02Qr8C5SNRHe3N5wUE0TEwtN6ZfPFi2fPtOwkiZkLn 9a5 x1TcQCM9R6PYcFuVi8f9tGESRORV4dGvCcfe5TF7o427wGmhEGZJwgRWUDBHL3caMobJnMJprvret7/z6PGT9eZMRGobnuEYIIl7mI3XnOj/Ep4CENJpf/44VHig+RALLTJu93EWgddxXE03CaNf5OJw34lCfTTuKWUzOn/05O15/5t/+vmh1lUaKPSm9dQfB08MJA5zU49mMCPAW2gLDOE3OdlEFK6GUEQbEztzCgoKYwoCC5AQeciZkwgS2pTNQMfgHEUk4rQvhVJO03Aourk4//Cjb24uHq825wRx1yZrOLbl4s1GKqca3X9m0Xd/oH9bI/Tqov2ICMXdfkBnTtytYONzlXvtp5vfu9cuZjQ3zylLzmfr9ZBl3l5fPv/MyDPkSC 3A4 l5iDgEFcQKTOUgQS+cgmMMNpwQcz+QczEsUnLBbmIATM4gys4CnnGUxeDSYVSd+kxu43/JzysN4PZd9rX/ywx9uHj8JMOdMIA3PnG9O/fj8u2w/Xh/YpB+o+/BGOaVEJ/2Apfr4wu6TG3E7iD3sCxwld9fN8pVjw3xar4hoGMb33nsvCV1+9knb7U+/yTEYoOlwwUzeckWpRxcCDGlyBGo8h2huQYpwALLAmhnU7GCJmanNmKklCbWoC/BNP8LdwVRVg+jDj77x+MlbNE4l0IS8R/D6/c7NlzVz+WLu0hQI59OK7aYSkI5ypNRC47tk74YT0o+MliTNCwOJGhdGCMcicukwLjirI2RO+l0fi9rTW0/TFwFxF/2RuOk4rtd5JNU0TZsPPjq7OP9Fks8++R0FZwaC3KwLx 8A9 xoU7m6MTz9yIOynH6bji2ZZzpC9gQQoMLAkCtFxCdqcFfwZmAQKzjePgSGZELMjjXs1Iv/Ht7/7+H/4gTWuk8fp6K8S6LxPnxQPZY2LgACPYoqdJMtiXGa7R0iyKk7BJAB5LushJIwQ4hlQHgJZMAG 4a8 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 2a9 /wT2+DUFkcAABP5n7Me6egU2k0UKBbk6O9soRnYfPeOi/AujNIMRvkGYIcpAiDH3f7GAfgjT600mz8BW1JBMjhEkAAQkgDm44SQc50P8haS6ZvkeByDQzffWtxxdnoyTJw5CHgQCLcJAL/9Gf/vhPfvyj1cX51f6Qx4mW2Vs4mYWZWzWv4Uph7TFoqSHMwk2Se+y3JLSEEHSrDSL6kNrJA25w40ZcY3GWy/1+ODu7nMt0tvnzv/qr3//+9x9/5W2NyKsxDXm1WpX9Yb 2a/ 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 3a9 +85u/+/ufXl5dV9M85EGyU1TV5fbPvkh4W23ii6B6aSEdB2Kfu7u4wK45nej2bypPPwka4OjR2jj6CIgsomVIH0FffO8HtO9wvGi9ePH8Yr0Osu328uxs9fF3v/3Wk8cffO29Dz94f8xsqhEeHBDOOSUkhJt6Ag8pI7FRJMLAjJbCqUYiwpIkd7oQS2ZRy/uyz5t1mMPtnXe/+tZbb23Wm//7Jz958fTpmPOQuNRDz9482WpbvlXPHQMElDgTJIxqGEso4g/++E8+/le//7VvfHSwQJJpHFPKYLBItzEwm3tDobpaCz9suYvCDA24hceT1ebRtHn7/NE3P/zGH/3xj//h5z/7 1a9 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 +a9 uqVKy/cuvUXP/qRY/aOC2bvfVUW7flJaGoLDmzBrA9+GEwsAGDd+Kp6fn4uIl3XxRiZ2br8m6YBjfau1mRlQxARMwLLlisWVRXxIYSiKFS1LMumXSBiURRbW1sHO1c+Pfrk449vf3HnD3V9TiKxaYloe2uLyS8Wi7OHp8+98K3vfO9Prt+4AQD7B0fXbjxxvqirqlBkdo4QY1TvXZJ6MNHyGKXrOtt21kWPoETkhv5DQowhWjM7M3vvJYI3+RsVC7+EYGtra/9g/+mnnwKA8xDf/eC3b7zxxltvvbU4P3VVBQJlUZLjxeIsdrEqp6khCMy3Yd8MbE0X0I+m6KdH67C4Fsg7k3AeandLd5N3g1kZyjOjQuiCaMAIpXdEKF29OF809dmVo4Nvf/s7L7744teeuLG7vdO1rWP0zqtIiN35yaPKUTmZGgjfdR3qsltWUMSiT+YQQicm0tQXNpjIO6eqEmOMy7 5a8 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 /a6 L77z33mtvvvn6m2+3TY3IrvBIDlRFokgEECLyVQGEIQQAcYd7+wXhvc8/r4piVvju/Exjp0y3vvHsN77+7K3nn9vZ3p6UZVl5w8+ZkAhKdn3fbhQAcExEaFJqIQTpCYPWGdPH+ei8eX2z1bb3864os5/JnufsBudcjJGYAMgmSZtuIHMfjtn/mu/XfoqWEDlmLyIfffTbV155tSzL6XRLFRlZWX3hvHeLxVld13XXqmpRFM4VGCRGdegQLRE3KjSqQllO6vr8ww8/un7txte//nUbRJdpf 6A1 tnjvvfPcj0236QLS5yyD0pPdIAHaWIIEngJAXddlWRa+8MyILI4VWGWQp9ah6Cr9KWIE6RoK3VZZzPb3D 3a3 n3366T/7wQ9f/vW//vajj+4/eIhMTL7wpcXVddfWZ+fomL0DAHf26AG13e50EtqmIv/N77xw67lvHO7vHu7tFt57x555Ni26riUicoWVr5nIsUMFG68YQDpRAAkgAQRFfFFY3aBAh4gm7KbKfSlz0BLLC9uIKBRtByR8o21be3Y96xcYe9XI/tiljqrhCSY5CHXszs4Wx8cnv/rlr8/Pz/d2DwgdAm1tTyfTcro1c47iabu1Nd 3a2 VaFKFTXnWCcb89KR129aOo6xqgAxAygbZCimj04Pvn1K699cf/Bi9/9blVVlpTaIEG7eO+89945JhB21sKmRMTYb2giGvTUgABpeA52C23baogROhBBACUEx+jQk2dmRO7aaGFODErYKylwVURROTstiJ44Orh2ePC9b3+76bpPP/v83ffeu3f3/gcffvTF3XuREJmmvgSmIEI2lZVVbj3//F/++Y+vHB3Up6eocWc+26oKVHCOp5Pq+OSBc7Q1nYUQ20UNiBIl9jgUCkIAjaQMjDZkEcAU9SEmWqPRWIGZEGHwAjzYAEAEEZTeG3JqjVosNMbgvUPEKMLsWZc4ge2bEELKykQic0GEqtA27fvvf/DBBx8cHx8fHBwQeu+L2WxeTnl7d7Kzs6Oqk2mB9+DTz+8AQFlt7eztTiazo4OjAuP52emDBw/OzxaLxcI2XIi9Rswnf/j8zr17B4eHzz77bNM0PVQF2OeHZCJkDLFPSmPsAJTZ90x26EcHymD/hkBBRXVSliISu6AhImLkPtJ16FGRCW2LA5ArlBBLx/X5w+3ZFADP6kXTdkVRevaRtORy/szTt579uqJ/eHz+m1dffeX11956792zxXkxqdg5VHA/+t53n7x65YkbN6qiUA27+9ueXdsuQts6Jgly8rAuCo+q58enTdMgUVEUvnDOs00RIiTTJEHAAAoSRdSxZyBBQUCFKCptiADqmJk8UZeadgDQEaJgJHSO0eBd9rai3hdt2zpXMGPTdoLgjGzdAaCWWIYYi7Ls2k5ERUEECYrYds4Xv7/9uzfeePPBgwdlOYkq27vzajJxzu3t7WzPq7pui9Lt7RweHVx79pnnbJqZcx6JQVS6rkCSNliGtlgsQghlWYYQNEYGlLZ75613rhwdiYh3TkS7riNUAGIEx8Bk9KpMpw2TeDUCgGPuQhgak/s526JCRIAgIAroiYlQYkQFkbaWBZLz3hMSszM9FFGdzbfapolRnfO+KEOUIKFwHpCCxBBbgW5/5v/6L/7DD7/3wr+9/8F7H3741jvv3HvwwKm45566uVe6oiiMug0gGhsGjRI7icLsTBje3DagGWQgm60exULMqE5pkAokIGBFRbZjGiOCRAUEAWQHCCrDlExr6UIEQhQMZGCUOmYACCGWZVnXrXOeiIii1cKICVAxADsExKZpm6YjNP2m6vjR2d72zie//+QX//zLe/fuz+fzajarJhN2bJpaqhJaCW24ce1GURQichbPiCjGGBd1H2Q0XRcCApbsGufM5NSLRYwRg4QYI8i/f/jB4eH+t771LWaW2JVFH+h4z0XBRBiM5SUDikU9w89MFHsnNpJTpe9y4L76S44Ue3yQAViBAQixjQIQQhMAgL0riiIKtiKipQChQyRCRMeIMZJGBCEV1hhBo4Qoulu6//RnL7307edvv/S9//rf/js0nZtWk+3ZlkErBkv2c7+HmiExQTbrCtcai3oSJcKIIraiiTjIYuU6RH36OqJrohlRUlVmzeR1kZkBtJd1MO3eqMzctcF7nzRLY4xN07z++uu3b9+ezmZ1XXNR+KIwwWDnqCxL5wkaabt6UZ+dnZ21bWs724JQEdnZ3laAaTcLIVTHjxRksVg09SKEDqOgaFQ5PT19//33n3nmmclkIiJFUaTrtBAHpVdnSzjmSlt43w+J45p9j2thklrF7J9kgDo0BDu3lhgnFcJ1aVoig/uJVOsuxBhmk8lz3/jG3/wN3r33hXkUiDEa0g4AW1tbCU7p6aSMMJo1nxWY7besXXxJurJfllOnRMYtOeIAAB3ESURBVC1oCl1nUE9fk0WyCLG/Yqb8E5bzUJxTMT14Jk9M1LZRJE4mlfeFBYhN0+zu7v7sn/7pjTffNHjAggaJsZhOd3d3rGriXImIJycntvY7OztF9qOqFDGKTE0TdDZFwtPTU40iIhAiKWjoAOmLL7545513Xnrppa2tLcsL0vqNzoDFf+k1htEys8SVA5P+NGdvS2t5tW3xoMLMIQRk8t4jURRJqYRtGhEZJVOoCowAjKAqQUTbNjz1xPXnn3vGxRjtGSX2tPc+wW19zr16TNcZ6TpEbv33rd5MWjyN/VV2IfR4FCIMAV26bbbS7oDspl44ZgcIEQRUmQjBqUMAh0AnJ6eqOplMvPfvvPvuv/7mN8g0n8yQ6OjoCB1XVXVweHD12rWmadrFOSKa2SjL8tq 1a/ P53LAE7733HgCg6yfTIOKeHOwf7H/66acESEShbmLbdedRNLZt++qrr16/fv3pp59OBz3teyS0AY2W35s+gZm4oarJIQmqZVlxghZS1rdUEglxNF2QmTnTDM4rbflCOKK2rQGxKB0iFd4F1cm0chKDxG6Q8bMDHUEjgoBGRCITcFw1UxuUQpdzrVcsf8IAgKmnAgxKyDFGZE57PJVzRo4mdxDMbK0g9rsIGAynqnbiP//881//+teLut7Z2ZlMp/P5/Nq 1a0 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 5a+ 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 9a3 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 1a9 +detbz1+5dnR6fu7cpCiKNoaoIAge0EoCCvqlT3g80oEGFzBIMn3pHCjrNOxHA6viIGxI66qyKdwb6cauyx3nsQzRSgv66gChcQ9hQqq99xpliEBpqas5vKlpWuN6d113fHw8m83u3v3itdfeePlffnn37gNmv729zcyHRwfXr1+5fuPq1tbU6GL5yD4LOyaTSV3XTdM0TeO8t/2XpIUT39cMm82KNjqbqho0wGzr7QYCI6lKjB 0A+ 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 5a8 8OO9913Uxhhg5mUfEDRvXlEoK54uicI6tW6Wu68RMtMTHOUYC55nJPvwMMO7szpp23nWNKrRtt1gs7t279+GHH+4f7FZVxTwcX8Qcz0/7b+Ny5I90BPtfNI9AL5X3SiEOdV2wb015aiJw2v92XWdEqLx4ZQ4CVxVobP+aaTV02XCxvJKRAG0isoDfvroHmhRMFyRRwq3VCwAmk4mqErL3ZVmW+/v7N2/e3Nqa1vX5YnGe/LdhRHGAGYxLbicvxmiOvGkaKwza5ZlfsHynd5yiIYSctmpctKqqrHI45PHaddHYQSG2oaeCEYDOZpPdve35fG4nfrFYfPbZZw8fPhzYzP1hMDqQfWnus+xdSecs0cgGfhgMgbMk7CihWz33fPBZFgXnFi5RLO01Lj8Bo0YLyOYfP44y+LqEzCiizMwGicSNoCFkeuvpYvrzKhqgm83Kg8P9vb3tauLB+Mk9VRhVQaKJw2pC0RNmZ40ZMca 2a5 cNazn7dqA4EJEn33d1DRP4NuXZaG3YA0VmSQ6oqnJ7e1siN013en7SDV1NSyOs0jeLAT6+OnSOCOUHD5U2rsUKbkv9jdjzTGCrG5UTUu9LOriXyN3r2gyj 5A7 SVSbAtf/8vlhCMYSR++jHrFNPErLyeb8GQ0g/n2/v7e3duHF1e2fGDG1bs+uDVkREcKbAk9fW7NCn6AkR264zYz4gwTyKXqlvC7TKOMoAUOZkBeumhX7qpclaLP1dVVVFUU6qbeeKpqvPFguzOvmzUhFdHc5y0SZYkkpWPUKeIGwsxeUvo6EGn9BugwoMFdDRGV3/my+dhTMOFNbiQQsXZTAtfdALeHkSnPYsMxel39vb2d3dnkxKU4FWEzMx1Z2+03JZXsrdgZ1si6HKsujp571vwmH5l3VIXc6/VBVNSVruH0X6GEJEiVAMXOz79qgoXOF9Xbfz+VzXtJ/W6VyXb4JUs022feN6XbT8o00zoMCICQ9IScioIpdzvB7fBYxGYeSRYDpMFouO4hcb42JdJokBZ6/c2dmZz+dXr16bTKqy8kiKJMwUY3SO0JRIYy9N7VzvIw0eSHiz+eAoYjwo+1LL/lPebw1Vgq. Research Activities The IFAC organizes every 3 year (triennial) conferences on marine systems including CAMS (Control Applications in Marine Systems) , MCMC (Manoeuvring and Control of Marine. Crafts). The scopes of these IFAC conferences on marine control systems are broad ranges from autopilot to dynamic positioning systems and various applications of control theories in control,