Trans RINA, Vol 157, Part A3, Intl J Maritime Eng, Jul-Sep 2015
3. Randomized anomaly tracks data 4. Divide normal tracks data into two groups using hold-out methods
5. Divide anomaly tracks data into two groups using hold-out methods
6. Combine first group of normal tracks data and first group anomaly tracks data, and randomized it for training process
7. BNs classification, get the accuracy of training data (memory test)
8. Combine second group of normal tracks data and second group anomaly tracks data, and randomize it for testing process
9. BNs Classification, get the accuracy of testing data (blind test)
The flowchart of the above design process is presented once again in Figure 2.
5. THE DATA
We use the AIS raw data of Port Klang from July to September 2013. The AIS raw data consist of 9,845 rows of data, including the vessel’s MMSI, status, speed, longitude, latitude, course, heading and timestamp (see Table 1). All information is obtained from the public website,
marinetrafic.com.
Below is an example of seven different vessel
movements’ data from the original AIS data. To retain the anonymity, some details are removed.
The AIS data were cleaned and separated into ‘tracks’ based on the MMSI. The above data consist of 367 tracks with 7 unique MMSI averaging at 1,400 rows each.
The vessel record data contains 8 variables, including the MMSI, status, speed, longitude, latitude, course, heading and timestamp.
6. RESULTS AND DISCUSSION Figure 3 displays
the scenario of vessel anomaly
behavior with the normal route of the Straits of Malacca with the longitude from 101.50 to 1020, and latitude from 2.250 to 2.750. As shown, the yellow circle indicates the normal speed of the vessel.
Figure 2. Scenario of vessel anomaly behavior with normal route
Here, we present the example of the vessel anomaly behavior in a speeding scenario. The variables that we used for training the model include the MMSI, speed, course, longitude and latitude. The spatial variable is also important in detecting the vessel anomaly behavior. For example, if the vessel moves with low speed and the location is near the port, it is considered as a normal behaviour. However, if the location is far from the port, it has the potential to be an anomaly behavior. On the other hand, the vessel also has the potential to be anomalous when moving on the waterway with speed exceeding the maximum speed.
Table 1. An example of Seven different vessel data from the original AIS data MMSI
Status Speed
47724**** 5 0 52501**** 9 222 53301**** 0 76 53301**** 0 96 53301**** 0 106 53313**** 5 0 53386**** 1 36
101.3038 101.3763 101.9443 101.1952 101.3227 101.3039 102.1907
246 269 304 29
227 164
LON LAT Course 2.951465 2.997293 2.274085 2.9433303
2.981503 46 2.951412 2.185028
Heading TIMESTAMP (UTC)
210 511 511 511 47 30
308
7/10/2013 9:20:00 PM 7/31/2013
5:15:00 AM 8/23/2013
9:22:00 AM 8/5/2013
5:03:00 PM 7/12/2013
8:26:00 AM 7/18/2013
6:32:00 AM 7/31/2013 9:44:00 PM
©2015: The Royal Institution of Naval Architects
A-149
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76