Defining Satellite AIS and First Pass Detection - exactEarth

Document Sample
Defining Satellite AIS and First Pass Detection - exactEarth Powered By Docstoc
					An exactEarth White Paper
November 2012

Satellite AIS and First Pass Detection
Executive Overview
As the world’s oceans occupy almost eighty percent of the earth’s surface, the ability to effectively
monitor and correctly identify vessels exhibiting behaviours of interest is a daunting challenge.
Additionally, with over 90,000 vessels navigating the world’s waterways, the amount of information
available was, until recently, limited to vessels navigating within the range of shore based AIS stations.
As a result, all ships transiting across the oceans went undetected, resulting in lack of visibility and very
limited knowledge on maritime traffic and activity beyond this shore based coverage.

With the advent of new space based sensors, namely Satellite AIS, the possibility of detecting and
identifying all vessels regardless of location has become a reality however it poses additional challenges.
Firstly, with the abundance of maritime traffic information readily available, the search for behaviours of
interest in the vast amounts of data must be examined in order to observe behavioural anomalies.

Secondly, combining information from multiple space-based sensors (Satellite AIS and Space based
Radar satellites) provides an opportunity to create a fused data set which can be used to identify vessels
of interest (both the behavior of interest such as an oil slick, and the probable identity of the vessel
through AIS), however this capability has yet to mature as Satellite AIS technologies have had generally
poor detection capability in dense shipping areas resulting in limited ability to identify the vessels.

This White Paper follows on from the previously released ‘Introduction to Satellite AIS’ White Paper and
discusses these challenges in more detail and how exactEarth’s use of Spectrum Decollision Processing
(SDP) can dramatically elevate the level of detection required to correctly identify and correlate vessels
with their behaviours as captured via radar and/or optical imagery satellites. Furthermore, a new term
will be introduced in this paper, First Pass Detection, and will discuss the rationale behind why First Pass
Detection is critical for proper identification and correlation of vessels.

Defining Satellite AIS and First Pass Detection
The Automatic Identification System (AIS) is a system that is used for ship to ship, ship to shore, and
shore to ship communications ranging in applications from collision avoidance to security. Terrestrial
AIS surveillance or monitoring systems have the benefit of continuous coverage and detection rates that
approach 100% close-in to shore. These systems also have the disadvantage of very limited range and
high cost per square mile covered. Satellite AIS (S-AIS) addresses both of these disadvantages as it is able
to provide complete global coverage as well as a low cost per square mile covered. However, to achieve
effective operational use of S-AIS data, and since AIS was never originally designed for satellite
reception, the capability must effectively mitigate the challenges presented by time slot collisions
inherent to the AIS system, which impedes the ability to detect a high percentage of ships via satellite.

An effective measurement of a Satellite AIS constellation’s ability to detect a high percentage of ships is
called First Pass Detection (FPD). FPD is the ability for a single satellite to maximize the reception of AIS
position report broadcasts from ships in order to satisfy operational maritime requirements. Given that
ship populations are not static it is of paramount importance to maximize the number of ships detected
during each and every satellite pass. A pass is defined as a subset of a complete satellite orbit over
which a given ship is within the satellite’s field of view (nominally a maximum of 10 minutes per pass), as
it is likely that when another satellite passes, any given ship may have moved in to, or out of, the
satellite field of view.

The deployment of a highly effective Satellite AIS constellation capable of excellent FPD faces many
challenges due to the fact that AIS was primarily intended for local sea-level transmission and reception.
The saturation of the satellite receiver due to the high amount of AIS message broadcasts, particularly in
highly dense shipping areas such as the Mediterranean and Baltic Seas, and the Straits of Malacca can
create a significant technical barrier to developing an accurate picture for maritime domain awareness
using satellite AIS only.

Recent advancements in satellite technology have led to the creation of higher detection performance
AIS satellites that are able to detect a significantly higher number of ships during a single satellite pass
than any other system, even in dense shipping regions of the world. This paper will compare the
message detection performance over a single satellite pass; with one satellite using the conventional
non-coherent receivers typically employed for on-board processing (OBP), and another satellite using
powerful coherent processing technology known as spectrum de-collision processing (SDP). In addition,
case studies are provided highlighting the importance of FPD for various operational maritime use cases.

Satellite AIS Processing Techniques and
First Pass Detection
The completeness of the maritime traffic domain for any given Area of Interest (AOI) can be defined as
the ability to detect ships and the latency for any given ship position report. Latency is defined as the
time interval between the reception of the ship’s position report on the satellite to delivery of that
report to the end user. Latency is impacted not only by the mechanics of satellite constellation (i.e. the
time it takes for the satellite to find and downlink data to an Earth Station), but also by the ability to
detect as many ships as possible within a single pass over the AOI. If multiple satellite passes are
required to detect all ships, the true latency is significantly higher than any S-AIS service with a high first
pass detection rate as it will take a substantial longer period of time to obtain a complete picture of the
AOI. Additionally, ships are moving targets therefore each satellite may have only a single opportunity to
detect the ship before it travels out of the AOI.

The likelihood of first pass detection can be impacted by a number of factors including ship density, AIS
message broadcast frequency, and in-band Radio Frequency (RF) interference. Understanding that
there are many factors that can impact detection, AIS time slot collisions and satellite receiver
saturation are, by far, the main factors impacting an AIS satellite’s FPD performance and the detection
methodology has the largest impact on mitigating both message collision and receiver saturation.

There are two detection methodologies supported by AIS satellite constellations currently in orbit:

On-board processing (OBP)
This processing mode essentially decodes AIS messages directly on the satellite using narrow band filters
and stores the messages on the spacecraft for later downlink to the nearest Earth Station. It does not
require any special processing and is effective in very low density areas, such as the middle of the Pacific
Ocean. However, the detection probability is low in areas where the satellite footprint (~5,000 km in
diameter) contains a ship density exceeding about 500 ships as it become likely that all of the AIS
message slots are being used by more than ship at a time. This effect results in a time slot collision and
it becomes much more difficult to successfully decode.

Statistical analysis has shown that the first pass detection performance for OBP in high dense areas is
much lower than spectrum de-collision processing over the same area. OBP suffers this poor detection
performance partially because of its inability to deal with these time slot collisions inherent to the AIS
system when received from space. With a low per pass probability of detection across high density ship
populations, effective operational capability is seldom achieved with S-AIS OBP only.

Spectrum de-collision processing (SDP)
SDP requires the capture of the AIS RF spectrum and processing of that spectrum using highly
specialized algorithms to successfully decode AIS reports. With SDP the first pass detection ability is
high even within high ship density areas, thus quickly achieving effective operational maritime
capability. Statistical analysis has shown that the improvement in FPD for this detection methodology in
highly dense shipping areas can result in significantly higher ship detection, number of position reports
received (for behavioral modeling and predictive algorithms), and reception of multipart (multi-slot)
messages containing static and voyage related information. If the AOI contains higher ship densities,
SDP is required to achieve FPD performance that will enable operational use of Satellite AIS to develop
an accurate picture of maritime traffic.

Comparison of varying methods of Satellite AIS reception techniques: OBP and
The figures below effectively demonstrate a measured and absolute detection performance
improvement based on analysis conducted using onboard AIS signal decoding with OBP and using AIS RF
spectrum capture with SDP. This analysis was conducted on October 23, 2012, using exactEarth satellites
operating in both modes.

Example Area: North-East United States

 OBP                                                  SDP
 OBP                                                  -1

Processing Mode                                  AIS Messages                           Vessels Detected

SDP                                                      9,321                                      1,969

OBP                                                        272                                        227

Applications Enabled by Higher Rates of
First Pass Detection
Ship Identification with Radar and Optical Imagery
The First Pass Detection performance realized through SDP processing of Satellite AIS data has proven
itself to be highly valuable for leveraging imaging gathered from radar satellites. SDP processing
guarantees the maximum first pass yield for ship detection, therefore is much more likely to properly
correlate ship positions with images contained within the radar capture. Once positional correlation is
achieved, SDP provides additional advantage because it is able to not only capture many position
reports that are key to effective predictive algorithms, but also much more likely to receive the static
and voyage information transmitted by ships. This information can then be used to correlate to other
ship registry information allowing the authorities to confirm the ship is in fact the ship it claims to be.
The voyage data can be used to cross check the position and course over ground to confirm that the ship
is reporting accurately.

High first pass detection performance is fundamentally important to fusion of radar/optical imagery
with satellite AIS and therefore identification of ships. The radar image provides virtually 100%
detection of large ships (by design) as an anonymous target, but identifying those ships and using that
information to identify a ship of interest can only be achieved quickly if the Satellite AIS data contains a
report from all (or most) ships captured in the image.

The number of AIS position reports acquired during the Satellite AIS pass are used in predictive
algorithms and to fully identify any given ship. The static data that identifies the IMO hull number and
ship name and callsign is contained in a multi-slot message, as discussed earlier, the most likely
detection of multi-slot messages is achieved using SDP. The accuracy of predictive algorithms is
dependent upon the amount and quality of the data received, the more AIS position reports the more
likely the predicted position will be.

Fusing Satellite AIS with radar and optical imagery can assist enormously in creating an accurate picture
of the maritime domain to provide near real-time data eliminating the need to wait for interpolation or
forecasting. It also helps to improve ship validation by implementing measurement tools with radar
images as well as improve accuracy for ship identifications.

Search and Rescue
Satellite AIS is an important tool for search and rescue efforts enabling authorities to identify ships
within proximity of a distress situation, track a drifting vessel, and enhance planning efficiency both for
routine and emergency situations. With effective first pass detection, authorities can identify the
majority of ships within proximity of a vessel in distress, radio those vessels and coordinate immediate
and effective rescue operations in an emergency situation. By examining the heading, course over
ground, and the navigational status contained in position reports, it is possible to determine if a vessel is
drifting and if the current and wind direction in the area is known, predict where the vessel will be once
search and rescue assets arrive. When coordinating routine patrols the traffic pattern in the area of
interest can be examined and this information can be used to increase the efficiency and effectiveness
of those routine patrols. Finally, in an emergency situation the richness of the data obtained through
SDP can be used to provide an accurate prediction of the location for the vessel in distress.

Unfortunately, search and rescue efforts often involve ships that are no longer broadcasting position or
have access to radio communications. In these situations the quantity of position reports received
through SDP can greatly increase the likelihood of determining the predicted positon of the vessel in
distress when the AIS broadcasts cease. Examining traffic patterns to establish common routes allows
for historical anaylsis to complete the real time picture, effectively identifying predicted behaviour to
coordinate SAR patrol craft for investigation.

Environmental and Safety Forensics
To protect offshore installations and reduce the risk of a maritime casualty and resulting marine
pollution, traffic exclusion zones and precautionary areas have been established. Precautionary areas
might envelop wind farms or oil and gas fields, and apply to vessels of a certain size. Vessels may be
prohibited from entering certain sea areas due to environmental concerns. Other areas may be
prohibited or tightly controlled due to ecological disasters, or to allow endangered marine species and
organisms to thrive or recover.

Satellite AIS provides accurate information on ship behavior and position such that determining, or
definitively concluding offenses such as pollution violations, route violations, or irresponsible navigation
for the purposes of prosecution or establishing innocence. While Satellite AIS can only report the
position information during a given pass, the quantity of those position reports can establish a pattern
of ship behavior where activity during the time when the ship is not being monitored can be fairly

The high first pass detection provided by SDP maximizes the probability that a ship of interest will be
identified when authorities are investigating incidents. In the event of a pollution violation it is
important to quickly determine all ships that may be responsible for the violation. In the event of
incursion violations, it is again important to maximize the probability of vessel detection either during
the actual incursion or with enough data to prove incursion. The number of position reports is
absolutely critical when evaluating behaviour of a ship, whether on a particular route or over time in an

area of interest. Proving route violations and irresponsible navigation requires accurate modeling which
can only be achieved through Satellite AIS and SDP, as there is no other practical wide area sensor over
the ocean areas. The richness of multiple position reports can provide the data necessary to accurately
predict and confirm behavior during coverage gaps.

Satellite AIS is an emerging technology that provides advantages across a wide variety of applications.
However, without high rates of first pass detection (FPD) provided by spectrum de-collision processing
(SDP), those applications are limited to a smaller sampling of the maritime traffic rather than providing a
more complete picture. After an in depth comparison analysis, onboard processing is adequate in low
density shipping areas of the world but as the number of ships utilizing AIS continues to increase, those
areas will become increasingly rare. Operational FPD must maximize the number of ships detected
during each and every pass, and only through SDP can the constellation effectively increase the
reception levels for all AIS messages.

First Pass Detection using SDP provides a complete and comprehensive view into shipping activity across
the globe regardless of ship density and proves to be the critical enabler for achieving true maritime
domain awareness.


Shared By: