V.M. Grinyak - Ph. D. (Eng.), Associate Professor, Far Eastern Federal University (FEFU) (Vladivostok). E-mail: viktor.grinyak@vvsu.ru A.S. Devyatisilny - Dr. Sc. (Eng.), Senior Research Scientist, Institute of automation and control processes (IACP) FEBRAS (Vladivostok). E-mail: devyatis@iacp.dvo.ru

Vessel traffic control is the part of a vehicle control science. Vessel control systems (VTS) is the main instrument of marine safety in restricted waters. VTSs are based on radar and AIS information about observed ships. Ship collision risk assessment and collision avoidance based on ships trajectory (coordinates, course, velocity) estimation and extrapolation. If the system solved the risk assessment of ships then alarm and some collision avoidance scenarios would generates. Ship trajectory extrapolation always is indefinite. It is need describe of risk level with some terms, such as "no dangerous", "low dangerous", "very dangerous" etc. So, it is let to simplify water situation solving by captains and VTS operators. The problem of marine vessel traffic control is discussed in this paper. Information system model for marine safety estimation (risk assessment) is watched. System defines some alarm levels: "green", "yellow" and "red". An alarm criterion is based on maneuver detector. If the ship is maneuvering its alarm level decreases ("yellow"). Mathematical model of position and velocity estimation and two neural fuzzy network (ANFIS) configuration for alarm generating are offered. The first neural fuzzy network is maneuver detector. It processes data of α-β filters with different numbers of measurements. Network outputs the ship maneuver level. The next (second) neural fuzzy network is alarm generator. It inputs maneuver level, time of closes point of approach (TCPA), and ships relative movement parameters. Maneuver level is described by two terms ? "maneuverable" and "constant". TCPA is described by three terms ? "little", "average" and "large". It is two parameters for relative motions of sips describe - angle between line of sight (LOS) and relative velocity vector and angle between LOS and sips domain borders direction. Its fraction is described by two terms ? "big" and "little". There are 12 roles in the second neural fuzzy network, its output is alarm level. Networks are learned by expert, no training sample for learning. Some results of experiments are shown, such as numerical experiments for typical ships traces and field experiments for Vladivostok port water area vessel traffic. There are near 20% "yellow" alarms and 80% "red" alarms. Its alarms are generated in all water area.

 

1. Grinjak V.M., Golovchenko B.S., Malko V.N. Raspoznavanie opasnykh situacijj sistemami upravlenija dvizheniem sudov // Transport: nauka, tekhnika, upravlenie. 2011. № 8. S. 42-45.
2. Grinjak V.M., Devjatisilnyjj A.S. Nechetkaja sistema raspoznavanija opasnogo sblizhenija sudov na morskikh akvatorijakh // Vestnik kompjuternykh i informacionnykh tekhnologijj. 2014. № 11. S. 36-42.
3. Tam Ch.K., Bucknall R., Greig A. Review of collision avoidance and path planning methods for ships in close range encounters // Journal of Navigation. 2009. V. 62, № 3. P. 455-476.
4. Burmaka A.I. Strategija raskhozhdenija sudov v situacii chrezmernogo sblizhenija // Vestnik gosudarstvennogo universiteta morskogo i rechnogo flota im. admirala S.O. Makarova. 2014. № 1. S. 20-22.
5. Wilson P.A., Harris C.J., Hong X. A Line of sight counteraction navigation algorithm for ship encounter collision avoidance // Journal of Navigation. 2003. V. 56. № 1. P. 111-121.
6. Bukatyjj V.M., Morozova S.JU. Netradicionnyjj metod vyjasnenija situacii sblizhenija sudov // EHkspluatacija morskogo transporta. 2012. № 1. S. 18-21.
7. Dorozhko V.M., Lebedeva A.N. EHkspertnye predstavlenija ob osnovnykh situacionnykh modeljakh kollektivnogo dvizhenija sudov // Problemy upravlenija. 2006. № 4. S. 43-49.
8. Vaskov A.S., Vaskov V.A., Mironenko A.A. Vzaimosvjaz zon navigacionnojj bezopasnosti sudna // Vestnik gosudarstvennogo morskogo universiteta im. admirala F.F. Ushakova. 2013. № 2. S. 18-21.
9. Farina A., Studer F. Cifrovaja obrabotka radiolokacionnojj informacii. Soprovozhdenie celejj. M.: Radio i svjaz. 1993.
10. Devjatisilnyjj A.S., Grinjak V.M. Nechetkoe opredelenie kharaktera dvizhenija pri mnogomodelnom soprovozhdenii traektorii sudna obzornojj RLS // Nejjrokompjutery: razrabotka, primenenie. 2013. № 6. S. 13-20.
11. SHtovba S.D. Proektirovanie nechetkikh sistem sredstvami MatLab. M.: Gorjachaja linija telekom. 2007. 288s.
12. Konoplev M.A. Primenenie apparata nechetkojj logiki dlja opredelenija urovnja opasnosti stolknovenija // EHkspluatacija morskogo transporta. 2009. № 2. S. 34-39.
13. Bobyr M.V. Vlijanie chisla pravil na obuchenie nechetko-logicheskojj sistemy // Vestnik kompjuternykh i informacionnykh tekhnologijj. 2014. № 11. S. 28-35.
14. Golovchenko B.S., Grinjak V.M. Informacionnaja sistema sbora dannykh o dvizhenii sudov na morskojj akvatorii // Vestnik gosudarstvennogo universiteta morskogo i rechnogo flota im. admirala S.O. Makarova. 2014. № 2. S. 156-162.