Beitrag 10

Ergebnisse im Projekt FAVECO - Antikollisions- und Strandungsverhütungssystem für schnelle Schiffe

Prof. Dr.-Ing. R. Müller
Hochschule Wismar, Fachbereich Seefahrt

  1. Introduction
  2. Problems
  3. Testing the application of FMCW-Radar technique in shipping
  4. Look Ahead Function
  5. Technical data of the tested Radar


Beim Einsatz schneller Schiffe hat sich in der Navigation bei hohen Annäherungsgeschwindigkeiten der Fahrzeuge das Impuls-Radar als alleiniger Sensor für die Kollisions- und Strandungsverhütung als nicht ausreichend erwiesen. Auch die benutzten Bedienoberflächen, wie auch insgesamt die Mensch-Maschine-Schnittstelle berücksichtigen nicht die oft sehr kurzen Reaktionszeiten für das Manövrieren bei diesen Fahrzeugen.

Im Projekt FAVECO wurden Anzeigen von Impuls- und eines neuartigen FMCW-Radars als gemeinsame Sensoren für die Kollisionsvermeidung versuchsweise auf der Schnellfähre "Delphin" installiert und ausgewertet. Angehängt wurde ein im Projekt entwickelter Bahnprädiktor für schnelle Schiffe. Auf der Basis der Sensordaten und ECDIS-Informationen wurden Ausweichbahnen gegenüber festen und bewegten Zielen erzeugt und zur Anzeige gebracht.

Eine Koppelung der Steuerausgaben des Bahnprädiktors an das Ship Control Center SCC eines Fahrzeuges wurde am Maritimen Simulationszentrum (MSCW) in Warnemünde untersucht und erprobt. Eine audiovisuelle Bedienoberfläche für die relativ schnellen Prozesse der Kollisions- und Strandungsverhütung wurde entworfen.

1. Introduction

High Speed Craft (HSC) has been gaining importance world-wide. Vessels with rates of up to 40 knots have become more common (e.g. London-Amsterdam, Holyhead-Dublin, Skagen-Larvik, Hirtshals-Kristiansand, Rostock-Trelleborg, Rostock-Gedser). Reducing the travelling time has become crucial for ferries in order to attract customers and remain competitive.

The use of HSC´s is not limited to ferry shipping. In the field of conventional navigation, feed time has also been increasing in importance. The first orders for fast ships for freight transport are to be expected in the near future.

Currently, fast ships are still equipped with conventional navigation systems. Due to the ship's own maneuvering potential, encountering targets can be regarded as motionless, and suitable avoidance actions are thus easy to determine. Conflicts of encountering HSC´s will become more frequent in the near future.

2. Problems

In an encountering situation between HSC´s, the tasks of target detection, the target's tracking and deduction of own maneuvers are concentrated into a short time. The technical support of traditional navigational equipment is currently insufficient for collision and grounding avoidance. Existing problems lie in the areas of automatic target acquisition and in the procedure of target tracking (including small targets). The available trial mode (see Diagram 1), based on ARPA procedures, is inadequate for own-ship track prediction. The handling of the trial function requires too much time and concentration of the officer on watch OOW. The use of the trial mode results in a variation of the own-ship´s speed vector for finding a first maneuver in a ship-ship conflict. An automatic generated complete track prediction including back to the previous track is not available.

However, if several fast vessels are running in one sea-area, the potential risk of collision increases enormously. Encounter situations between vessels with approach velocities of up to 140 km/h can occur. The navigating officer must select and interpret information in a limited amount of time after having received the data. He is under enormous time pressure which could lead to a false decision (due to the stress that may develop) during the planning of avoidance maneuvers.

The project FAVECO´s goal, in which the DaimlerChrysler Aerospace Ulm, TT-Line Travemünde and the Wismar University are involved, is to improve the mentioned technical inefficiencies. The project is sponsored by the Federal Ministry for Education, Science, Research, Technology and Building and supported by German Lloyd.

In order to guarantee safety and efficiency of the future maritime traffic, a reorientation is necessary for both the selection of navigational equipment and in the control concept.

The research and development project "FAVECO" was launched to improve safety in two aspects:

  • Look Ahead Function for decision support in collision avoidance
    • analysis of encounter situations
    • graphic indication of the area available to perform maneuvers
    • simulation of the developing encounter situation with dissolution of conflicts
  • Improvement of the radar-technical navigation sensors
  • testing the application of FMCW radar technique in shipping
    (detection of poorly reflecting targets in the close-ahead-area)

3. Testing the application of FMCW-Radar technique in shipping

In co-operation with the project partners DaimlerChrysler Aerospace and TT-Line, the possibilities of using a frequency scanning FMCW-Radar (FMCW=Frequency Modulated Continuous Wave) technique with a high scanning rate in shipping are examined. A first testing aboard HSC "Delphin" was carried out in September '98 and June '99 (see diagram 4). On the basis of recorded radar-raw-data the tracking algorithms are presently being optimized. The goal set to be reached in radar development is the automatic detection of targets, including small targets, and the implementation of a stable target tracking.


An improved target detection of small targets and a qualified target tracking, especially in short distances, are the prerequisites for a meaningful track prediction during a ship-ship confrontation. A future overlapping of impulse and frequency scanning FMCW radar's would be a sensor technical pre-requirement for an advisory aid for collision avoidance of HSC´s.

4. Look Ahead Function

On the basis of radar data, the analysis of traffic conditions is executed and - in the case of an existing collision danger - possible avoidance maneuvers are calculated. These are characterized by observing both the laws (COLREG 72) and the rules of good seamanship. The continuous display of a spectrum of complete avoidance tracks enables the navigating officer to make maneuver decisions safer, faster and more competent.

Schematic Structure Of The Look Ahead Function

The integration of the Look Ahead Function into the bridge equipment requires no additional monitor to be installed. The relevant functions of the control display are to be merged into existing ARPA technique (e.g. as substantial improvement of the trial mode offered at present).

Applied radar data

The analysis of the surrounding situation of the own-ship is executed using the following radar data, which is supplied by ARPA systems as appropriate standard NMEA data records (RAOSD, RATTM):


  • Course
  • Speed

Acquired Targets

  • Bearing
  • Range
  • Course
  • Speed

Further parameters necessary for the situation evaluation, which accord with COLREG'72, are derived from this data. Thus, the quality of the avoidance tracks produced by the Look Ahead Function depends solely on the accuracy of the radar information. The radar data can be simulated by a situation track generator for test purposes. This generator was developed in order to produce an immense number of encounter scenarios (randomly), to simulate the movement of the targets in real time and to supply the necessary NMEA data records for the interface.

Additional input modes

If the OOW has information which has not been detected by the sensors, he can actively intervene in the situation analysis by inputting additional data, e.g.:

  • Visibility conditions
    default: good visibility;
  • Status according to Rule 18 COLREG '72 (ownship and targets)
    input only with good visibility conditions;
    default: power-driven;

Development of collision avoidance tracks

The generation of collision avoidance tracks is executed by applying mathematical and computer methods based on artificial intelligence.

The algorithms analyses an encountering situation (if a risk of collision exists in accordance with COLREG 72) and produce three solutions, which correspond to different nautical strategies for collision avoidance:

  • Strategy 1: Course alterations prior to speed alterations (see diagram 3a)

    In this situation an avoidance track is favored, whose introduction is characterized by a preferred course alteration. After dissolution of conflict(s), the return to the origin track is included. In open waters this maneuver corresponds to good seamanship.

  • Strategy 2: Speed alterations prior to course alterations (see diagram 3b)

    This strategy produces a maneuver track, which realizes the minimum passing distance to all targets with preference given to speed alterations. In several cases the exclusive use of speed alterations is not successful (e.g. targets on opposite courses). Additional course alterations are considered for track generation (combined maneuvers).
    Especially in restricted waters, the navigating officer will include the results of this strategy into his decision.

  • Strategy 3: Remaining passive as long as possible (see diagram 3c)

    The avoidance track generated following strategy 3 shows the last possible maneuver sequence in order to pass all targets at a safe distance. However, this proposed maneuver should not be mistaken under any circumstances for the compulsory action valid for a stand-on vessel (Rule 17 (b) COLREG).
    Contrary to strategy 1, this maneuver does not often correspond to good seamanship. The suggested avoidance action should be considered as the worst possibility of all existing solutions (Rule 8(a) COLREG), and due to the long period of passivity very hard maneuvers are necessary.
    For this reason, the solutions of strategy 3 are not to be interpreted as advisable.
    Instead of remaining passive, the comparison of tracks 1/2 and 3 gives the navigating officer a general view of the decreasing maneuver quality. Furthermore, the time fund and included maneuver alternatives are indicated.

In encounter situations with existing course keeping obligation, the navigating officer is informed accordingly, tracks are not offered. As soon as it becomes apparent that the give-way vessel is not taking appropriate action in compliance with COLREG '72, an acoustic message is given and avoidance maneuvers are generated and displayed.

Validity period of tracks

For generating avoidance tracks, the simulated maneuver start is extrapolated one minute into the future. Thus, the navigating officer has sufficient time for interpretation and inclusion of the displayed solutions into his avoidance decision.
The remaining validity period is indicated by a time beam counting down the 60 seconds. At lapse of the time interval or in case of intermediate occurrence of relevant data changes, the tracks are recalculated and displayed accordingly.

Real-time behavior

During operation the received target data is analyzed every 20 seconds. In case of target movement alterations, it is checked whether the avoidance tracks displayed are still valid or whether a complete re-valuation is necessary.
Relevant interventions of the navigating officer (e.g. modifications of the view conditions or the minimum closest approach) always cause an immediate new analysis of the encounter situation.

Dynamic display of generated maneuvers

There is the possibility to graphically simulate the future development of the encounter situation by selecting a track alternative. In absolute vector presentation the own-ship-track is built up step by step, while the target vectors move over the display accordingly.
In this way, the future passage steps with additional dissolution's of conflicts are well illustrated, the navigating officer receives very important information for planning avoidance maneuvers.

Technical data of the tested Radar

(modifications possible)
PrincipleFMCW + 2-D frequency scanning
Center frequency34,3 GHz
RF bandwidth / video bandwidth1.8 GHz / 600 kHz
Transmission power (effective)0.5 W
Antenna Beamwidth (azimuth)
Range encoding6.6 m
Scanning angle (azimuth)40°
Update frequency15 Hz
Instrumented Range20 m - 3.5 km
Detection of humans in calm seaup to 3 km
RF-unit dimensions W*H*D / weight80 * 44 * 18 cm3 / 15 kg (approx.)
Power supply RF-UNIT230 V~ or 28 V= / 100 W (approx.)
Power supply Processing230 V~ or 28 V= / 300 W (approx.)