DSpace university logo mark
詳細検索
Japanese | English 

NAOSITE : Nagasaki University's Academic Output SITE > 080 水産学部 > 080 紀要 > 長崎大学水産学部研究報告 > 第58号 >

水中物体追跡装置の開発に関する研究


ファイル 記述 サイズフォーマット
suisan58_001.pdf18.21 MBAdobe PDF本文ファイル

タイトル: 水中物体追跡装置の開発に関する研究
その他のタイトル: Studies on the Development of the Tracking System for the Underwater Moving Object
著者: 中根, 重勝
著者(別表記) : Nakane, Sigekatsu
発行日: 1985年11月
出版者: 長崎大学水産学部
引用: 長崎大学水産学部研究報告, v.58, pp.1-80; 1985
抄録: Since various electronic navigation systems, have been developed as measuring instruments of the position at sea, the position measurements can be automatically made at all times regardless of the weather and in all the space of the world. These systemes provide not only the knowledge of the right location but also very usefull pieces of information about the fishing industry, the marine survey and the exploitation of the ocean. Under these circumstances, new systems have been developed by the requirements for extension of application and high accuracy, together with the improvement of the instruments. The purpose of this study is to develope a simple and convenient system for tracking the underwater object and for recording the movement of its location continuously. As the electronic navigation system makes automatic location possible, location can be established at all times and in a wide area on the sea. The position of this object measured on the ship strongly depends upon the accuracy of the ship loca-tion also determined by the navigation system. Therefore, various locating systems were examined on their systematic and accidental errors. In addition, the accracy of position, namely its allowable error which is demanded in the fishing industry was discussed for each individual type of industry. 1. Accuracy of location in the fishing industry The accuracy of location desired in the fishing industry can be classified under two large groups. One is the case of nektonic fishes and the other is that of benthonic fishes. In the former case, a rather crude accuracy of about a nautical mile (N. M.) can be permitted. On the contrary, a systematic error less than 0.5 N. M. and then a very small accidental error is demanded in the latter case so as to obtain high reproducibility. Especially, a minute accuracy is desired for the trawl fishing and the crab basket fishing, because the trawl-net usually go up and down in the same region and the basket of crab must be deposited like a pin point. In marine research, an error of 1 N. M. is generally good enough for a marine survey, but the systematic error within 0.1 N. M. and high reproducibility are needed for setting an artificial fishing reef. A severe accuracy of 0.1 N. M. also desired in the research of the behavior and ecology of fishes. 2. Valuation of various navigation systems The accuracy of astronavigation in use so far is the same as those of Loran-A and the Omega systems. However, this system takes a long time to process data and cannot be used on cloudy days even in the daytime. In addition to this, its accuracy depends on the skill of an observer. Although the radar system and the early Loran-A and C systems had similar disadvantages, these systems have made continuous measurements possible with the aid of a computer-assisted automated receiver. The accidental error of the Decca navigator system, which is recognized as the most excellent one, is as small as 0.01 to 0.05 N. M., so that its reproducibility is high. In spite of this, it shoud be kept in mind that its systematic error varies in a narraw region near the coast because of influence of topography on the phase of radio wave. The Omega system covering all the world lacks precision, for its systematic error ranges from 2 to 5 N. M. The intermediate system between the above two is Loran-A, for which a mean error of 1.0 N. M. is estimated. However, improved Loran-C has an excellent reproducibility because of a small error of 0.1 to 0.5 N. M. The satellite navigation system having two transmitting frequencies (NNSS) gives the accurate position in a systematic error of 10 m and an accidental error of several dozen m, although it has a disadvantage that the time interval till the next measurement reach about 1 or 2 hours by the latitude of the location. The Decca system and Loran-C are most adequate for the purpose of the detection and tracking of underwater objects, because these systems provide a minute fix accuracy with a small error. The errors acompanied with these systems were examined on the land base point and experimental region of the sea. As the result, the errors range from 0.01 to 0.03 N. M., so that the position can be determined within an error of less than 0.05 N. M. by compensating for the systematic error of 0.33 N. M. for the Decca and 0.15 N. M. for Loran-C. 3. Construction of the tracking system and its problems (1) A transmitter-receiver unit with a transponder was applied in this study. Microphone A was attanched under water alongside of the ship and microphone B was situated at a distance of 100 m from the stern. The distance between microphone B and the stern was kept constant in order to form a base line between the two microphones. The length of this base line was measured by supersonic waves and its direction was adjusted to the ship's head. The location of the transponder C was determined from the distances of three sides AB, AC and CB. The latitude and longitude of this position was evaluated on the basis of the ship's position sensor. These data were recorded on the printer, and was tracked in detail and continuously by the aid of a plotter. (2) The direction of the base line agreed with the direction of the ship's head when the ship was moving straight. Under altering course, however, these directions were different because the towing rope extending from the stern to microphone B was in a curved line. Hence, the measured values of position C had large errors in the time from the beginning to the end of a altering. Since in a state like this the duration time of a altering depended on the velocity of the ship, and of the wind, and the wind direction, etc. We measured how long the rope continued to be curved when we altered 45° and 90° at a speed of 2 knots in the case where it was calm and where there was 7 m/s wind. In conclusion, the time could be successfully cut down provided that the altering course was perform ed about 30% beyond the scheduled angle and then taken back to the intended direction. The duration time was about 2 minutes for the altering of 45° and about 3 minutes for 90°. In addition, the speedup of the ship was also effective in reducing the duration time. (3) The direction of the base line varied corresponding to the yawing of the ship's head. However, the ship's head could be treated as constant despite the yawing, so long as the towing rope was kept in a straight line. While the ship was moving straight at a constant speed of 2 knots by an automatic steering, the angle between the point B and the fore and aft line of the ship was measured with an interval of 30 seconds by the repeater of gyrocompass. The amplitude of the yawing angle in calm condition was about 2° and its period was about 1.5 to 2 minutes. Since the ship was drifted to the leeward by a transverse wind with a speed of 7 m/s, the towing rope was bent windward from the fixed point on the stern. Although the mean deviation angle was about 4° in the case of a due transverse wind, its standard deviation was smaller than in the case of a head wind, and the ship's head was kept slightly steady direction. On the whole, the yawing ranged from 2 to 3° and it was safe to say the rope continued to be in nearly straight line. Therefore, ±2 to ±3° of play was allowed in the direction of ship's head for computation of the transponder position. In the case of a yawing angle greater than this value, the direction of the base line was altered and the computed position had fairly large error. 4. Accuracy of the measured distance and position (1) Along the wharf the base line between microphones A and B was fixed, and the boat was moved several hundred meters from the wharf with spaces of about 100 m, from which transponder C was hung underwater. The boat was an-chored for the measurement each time. What we got was the calculated distance that was determined by the triangulation from the base line and its included angles. We also measured distance on the basis of the method described above. The differece between the calculated distance and the distance by actual measurement was compared and discussed. The mean values were 1.2 m for AC line and 0.1 m for BC line respectively and the standard deviations were ±4.6 m and ±4.7 m respectively. These values of standard deviation were only 1.4% of the full length. The standard deviation of the difference was also estimated at every 100 m distance of AC or BC. These values fell within the range from ±2.6 m to ±5.4 m, that is, they were rather stable, hardly depending on the full length. Therefore, the ratio of the standard deviation to the full length had best be classified at 100 m or 200 m intervals of the full length. (2) The deviation of the measured position was 0.02 N. M. on the average, and only 0.04 N. M. even in the case of longer than 500 m. As the deviations of the horizontal axis and the vertical axis components were 23.9 m and 11.1 m respectively, they formed an error ellipse with the major semiaxis on the horizontal axis. The accuracy of position C was simulated by changing the angle ACB and the distance between AC and CB. As a result, we found this angle should be larger than 10° in order to obtain a high accuracy, that is, the distance to the underwater object should be within the range of 2~4 times longer than the base line. We also found that the difference in length between AC and CB should be less than one-half of the base line, that is, the underwater object should be nearly on the perpendicular bisector of the base line. The farther point C gets, the more important this formation becomes. 5. Experiments in tracking (1) While moving the research vessel with the transponder fixed, we measured and compared the distances between AC and CB on the system, and at the same time measured the angles ACB and CAB with the sextants. The result showed an error distribution of point C scattered within about 0.1 N. M., although point C had been expected to be fixed. The reason was that the location of ship was determined on the Decca system and an accidental error of 0.01~0.02 N. M. was attributed to this equipment. Another error was a systematic error due to the change of the ship's head. (2) Another experiment similar to the above was conducted; the motorboat with transponder C under the water moved in almost parallel with the ship, and, as in the former experiment, we made use of the measuring system and the sextants and we compared the results of both. The ratio of the deviation length to the full distance was nearly equal to the above case on the whole, although the values were somewhat larger in a small range of the full distance. The mean deviation was 19.8±17.1 m (less than 0.02 N. M.), which corresponded to the accidental error of the Decca system. (3) The tracks both of the ship in motion and the underwater object, which were plotted on a chart, were moor smooth by aid of the Decca system as a sensor of the ship's position than by aid of the Loran-C system. This was due to the dif-ference of their accidental errors. As the accidental error of the Decca system is smaller, the track of the underwater object determined on this system was closer to the real one. Since the line between the ship and the underwater object showed both the relative direction and the distance, we could tell how good the results were by observing its changing tendencies. It can be confirmed from the experiments described above that the position of the underwater object is successfully tracked by the present method with accuracy comparable to the accidental errors of the Decca or Loran-C systems. Since the present study is restricted to the measurements of the object moving horizontally, it is necessary to make a further study of the measurements of the depth of the object, the choice of the length of the base line and the measuring techniques in an altering condition in the future. / 本システムはトランスポンダを用いた応答送信方式によるトラッキング装置と船位センサとして電波航法装置を組合せたものであるが,最大の特徴は送受波器の1個を曳航することにより長い基準線を採用したことである。そして水中物体の移動を追跡しながら,船位を基準として相対位置を記録,図示する。その位置の精度について,各種の実験を行なった。トランスポンダと測距装置の両者を固定して,測距精度と測位精度について評価を行なった。偏差の平均値は5m以下,その標準偏差は4.7mで,トランスポンダまでの距離が増大しても偏差の変動幅はあまり変化せず,ほぼ一定している。実測位置の偏位は500m以上でも0.04海里にすぎず,平均で0.02海里以下である。さらにトランスポンダのみを固定した場合と,両者をほぼ平行に航走しながらの追跡測定の実験を行なった。固定したトランスポンダの位置には船位の誤差が含まれるから,その位置のプロットは1点に集中しないが,システムによる測位と六分儀による測位の差は大分部が25m以下にすぎなかった。両者航走中の距離の偏差や位置の偏位は,固定点での結果とほぼ同じかやや小さくなっている。偏位の平均値は19.8±17.1m(0.02海里)で,この値はデッカ位置の不定誤差に相当する。実際の追跡では,船位センサとしてデッカシステムを用いた場合の方が,この不定誤差が少ないからなめらかな航跡となり,水中物体の実際の移動状況に近い形で描かれている。船位からの水中物体の方位線は船首方向を基準とするから,基準線の方向と船首方向(ジィイロコース)のずれが大きくなれば,隣合う方位線が交叉する。また,実測距離の偏差が大きくて著しい偏位を生じたときも同様に方位線が交叉する。以上の各実験の結果から,本システムの測位精度は,測定距離の誤差のほか,船位センサとして用いるデッカやロランCシステムの不定誤差にも左右される。たとえ不定誤差の補正ができても船位センサの最小単位以下の測定位置の偏位は除去できない。しかし方位線の変化状況から,水中物体の移動状況や測定位置の良否の推定は可能である。従って水中物体を広範囲にわたって自由に追跡するという所期の目的を充分に達し得るといえる。なお,精度向上のためには距離ACのC点における交角θが10°以 下にならぬように,すなわち,基準線の2~4倍程度の範囲で,水中物体の位置が基準線の垂直二等分線上の付近にプロットされるように操船しながら追跡すべきである。
URI: http://hdl.handle.net/10069/30368
ISSN: 05471427
資料タイプ: Departmental Bulletin Paper
原稿種類: publisher
出現コレクション:第58号

引用URI : http://hdl.handle.net/10069/30368

このリポジトリに保管されている文献はすべて著作権により保護されています。
印刷やダウンロード等データの複製は、調査研究・教育または学習を目的とする場合に限定されます。

 

Valid XHTML 1.0! Copyright © 2006-2015 長崎大学附属図書館 - お問い合わせ Powerd by DSpace