무선인터넷 기반의 무인선박 제어에 관한 연구

Abstract

Unmanned system has been introduced to many fields. In the beginning, the majority of the system consisted of simple and repetitive automation technology that can perform tasks that are difficult or dangerous for human. Recently, however, other systems such as unmanned car, aircraft, submarine, ship have been researched and developed thanks to the advancement of technology. Google, the world's largest Internet searching company, has developed unmanned car and succeeded in test running. As can be seen here, unmanned system technology is ubiquitous in our reality.

So far, researches on unmanned system have been mostly performed at national level for military purposes or by some universities due to high development cost. However, the fact that such researches have been performed by commercial enterprises such as Google implies that the unmanned system technology will be commercialized before long.

Meanwhile, researches on USV are behind those on other unmanned systems, and Korea is no exception. However, surveys on market forecasts of USV indicate that such researches are urgent. First, American market forecast agency, Visiongain, estimated in 2008 that the market size of USV in 2009 will amount $160 million. Secondly, Moire Inc. estimated in 2003 that the market size in 2011 will be about 300 ships and $1.8 billion. Finally, the most recent survey performed by Research And Markets in 2010 estimated that the market size for the next decade (2010-2019) will range from 1,144 ships ($2.3 billion) to 1,870 ships ($3.8 billion). Altogether, this has been the background for our research on USV.

When USV carries out its mission at sea, every action including status check is taken by radio communication. Thus, selecting an appropriate means of radio communication is a crucial step. So far, VHF and/or UHF communication modem have been used for controlling USV, which need a separate operation console. Also, their communication range is greatly affected by antenna height, which decreases the mobility of control station and increases the development cost.

Thus, in this study, I would like to develop 'wireless Internet-based USV' that can communicate without such operation console and used for surveillance of major shore facilities and observation of marine environment. Control station will move the USV connected via wireless Internet to desired position. At the same time, the ship will transmit the data gathered by equipped camera and sensors to the station to enable real-time monitoring.

In order to develop such wireless Internet-based USV, we have applied wireless Internet service that has a globally established standard for communication and commercially available. Wireless Internet services that are currently available in Korea are WLAN with IEEE 802.11 b, g, n standards, WiBro (Mobile WiMax) with IEEE 802.16e standard and HSDPA that uses 3.5-th generation mobile radio communication network. WLAN is theoretically very fast (54Mbps) but has a very small coverage of 100m, and the lack of hand-over feature is its major weakness. WiBro has hand-over feature and can be used while moving, but its theoretical coverage is still limited to 1km. On the other hand, HSDPA has rather slow (theoretical) download speed of 14.4Mbps but a wide coverage as it uses mobile radio communication network.

Meanwhile, in order to apply HSPDA wireless to USV, it is necessary to figure out its actual Internet speed and service coverage at sea. For this, I have performed quality test of HSDPA wireless Internet during anchoring and navigation. As a result, the average download speed and average upload speed were 4.87KB/sec and 6.05KB/sec, respectively. Also, ping test result showed that both download and upload were free of data delay or loss, indicating that the data transmission was slow but very stable. Another interesting fact was that, although I had expected before test that Internet speed would depend on distance, the actual speed of HSDPA did not differ very much depending on distance.

It is important to note that upload speed is more important than download speed in operating USV. The data sent from control station to USV are text data that do not exceed 1KB, so the average download speed of 4.87KB is more than enough. However, data from various sensors such as GPS and camera image should be transmitted from the ship to the station, and the camera image data far exceed the average upload speed of 6.05KB. Thus, when designing communication program for USV, image size and transmission interval of image data should be considered. In this study, I have set the image size to be 105 x 75 pixels and the interval to be 3 sec. to resolve this problem.

Although I could verify the potential of HSDPA wireless Internet in controlling USV, it is thought to be more or less slow to be commercialized. However, once the 4th-generation mobile communication, LTE (Long Term Evolution) is commercialized, theoretical speed of HSDPA will increase by about 12 times. Moreover, the Ministry of Land, Transport and Maritime Affairs as well as 3 major telecom companies have been installing 45 mobile communication repeaters in coastal areas, islands and lighthouses since 2008, and have made an agreement that they will further install repeaters in 49 lighthouses until 2013 so that mobile phones can be used within distance of 30~50km. This will, in turn, increase the coverage of HSDPA. Thus, control of USV by HSDPA is thought to be practical in Korea's territorial waters.

Another feature of this study, along with the use of wireless Internet, is autonomous navigation. The goal of USV is not only to protect people from marine danger but also to save cost. Thus, USV without autonomous navigation feature does not have much economic profit compared to manned ship, as it should be monitored and controlled by someone at the control station. In this study, autonomous navigation is that the control station sets the waypoints of the ship through Internet, compares the ship position gathered by GPS and true course obtained from waypoint to heading gathered by heading sensor and enter it to controller, which will then autonomously adjust the rudder angle and navigate toward the waypoint.

In this study, in order to implement autonomous navigation, I have assumed the unmanned ship as discrete system and used ARX model (Autoregressive model with eXternal input) as linear I/O model of the discrete system. For estimating parameters of ARX model, I have used rudder angle () and turning circle angular speed () data that were gathered from the ship's turning circle test. Also, digital PID controller was applied to discrete system for autonomous navigation of the ship. Designing a digital PID controller requires (Proportional Gain), (Integral Time) and (Derivative Time), which were determined by Z-N tuning method. The wireless Internet-based USV was then successfully verified by actual tests at sea.