💻 computer science

High-Resolution Mapping of Port Dynamics from Open-Access AIS Data in Tokyo Bay

This study demonstrates that open-access AIS data can be used to reconstruct high-resolution port dynamics in Tokyo Bay, revealing precise berth locations, quantifying vessel traffic trends toward fewer but larger ships, and even identifying passive receiver stations through radio shadows.

Original authors: Moritz Hütten

Published 2026-01-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Moritz Hütten

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine Tokyo Bay as a giant, bustling highway for ships, where thousands of vessels zoom in and out every day to deliver goods, pick up passengers, or just wait their turn. This paper is like a high-tech traffic report that uses a clever trick to count every car on that highway, figure out exactly where they park, and even guess where the traffic cameras are hiding.

Here is the story of what the researchers found, explained simply:

1. The "Digital Flashlight" (AIS Data)

Ships carry a device called AIS (Automatic Identification System) that acts like a digital flashlight, constantly shouting out their location, speed, and identity to anyone listening. Usually, we think of this as a safety tool so ships don't crash into each other. But this paper shows that by listening to these "shouts" from the shore, we can build a super-detailed map of exactly what the port is doing.

The researchers listened to these signals for three months in 2024. Because the signals are free and open to the public, they could build a map with a resolution of about 30 meters (roughly the length of a bus). That's like looking at a city map where you can see individual parking spots, not just the whole neighborhood.

2. The "Ghost Ships" Problem

There was a tricky problem: Sometimes a ship stops talking.

Scenario A: The ship left the bay.
Scenario B: The ship is still there, but it turned off its "flashlight" (AIS) while waiting at a dock, or the signal got blocked by tall buildings.

If you just count the talking ships, you might think a ship disappeared when it was actually just hiding. The researchers created a smart "wait-and-see" rule. They said, "If a ship stops talking for a short time, we'll assume it's still there. If it stops talking for a very long time (like 48 hours), then we'll assume it left." This helped them distinguish between a ship that actually left and one that was just taking a nap.

3. The Big Finding: Fewer Bigger Ships

When they counted the traffic, they found a clear trend: The ships are getting bigger, but there are fewer of them.

The Analogy: Imagine a highway where, ten years ago, you saw 100 small delivery vans. Now, you see only 20 massive semi-trucks.
The Data: The average ship entering Tokyo Bay is now huge (over 11,000 tons). This confirms a global trend where shipping companies are using giant vessels to move more cargo at once, rather than many small boats.

4. The "Radio Shadow" Detective Work

This is the most creative part of the paper. In a city like Tokyo, tall skyscrapers block radio signals, creating "shadows" where the signal disappears.

The Metaphor: Imagine you are standing in a park, and you notice that the sound of a bird chirping suddenly stops whenever you look toward a specific tall building. By mapping exactly where the sound cuts off, you can figure out exactly where the bird is standing.
The Result: The researchers did the reverse. They looked at where the ships' signals were being blocked by buildings. By tracing these "shadow lines" back to their source, they were able to pinpoint the exact location of the listening stations (the antennas on the shore) with incredible accuracy—down to within a few hundred meters.
The Twist: Some of these listening stations are supposed to be anonymous or private. The paper reveals that in a dense city, you can't really hide a listening antenna because the city itself betrays its location through the shadows it casts on the ships' signals.

5. What They Mapped

Using this method, the team successfully:

Counted the average number of ships in the bay at any given time (about 381 total, with 35 moving and 346 waiting).
Identified 161 specific docking spots (berths) where ships park.
Mapped out the "lanes" ships use to avoid crashing, showing exactly where the fast ferries go versus the slow cargo ships.

Summary

In short, this paper proves that by listening to the "digital chatter" of ships, we can create a live, high-definition movie of port activity. It shows that Tokyo Bay is seeing a shift toward giant ships, and it accidentally (but cleverly) revealed the secret locations of the antennas listening to them, proving that in a crowded city, even a "silent" radio receiver leaves a shadow that can be traced.

Technical Summary: High-Resolution Mapping of Port Dynamics from Open-Access AIS Data in Tokyo Bay

Problem Statement
Tokyo Bay is one of the world's most congested maritime zones, handling approximately 20% of Japan's trade volume by weight and 30% by value. Despite its economic significance, scientific literature analyzing maritime activity in this region at high resolution remains sparse. Existing studies often rely on basic explorations of Automatic Identification System (AIS) data or satellite imagery (SAR), which can suffer from radar shadowing, false positives, and a lack of auxiliary data (e.g., vessel identity). Furthermore, while open-access AIS data is increasingly available, its utility for reconstructing port activities is complicated by uneven terrestrial receiver coverage, signal occlusion in dense urban environments, and the difficulty of distinguishing between vessels that have temporarily left the region of interest (ROI) and those that are moored with their transceivers switched off.

Methodology
The study analyzes 91 days of open-access AIS data (August 29 to October 27, 2024) from Tokyo Bay, collected via a terrestrial receiver network. The methodology involves three primary technical components:

Data Processing and Status Classification:
- The authors applied a cleaning algorithm to filter erroneous messages and classify vessel phases as moving or stationary.
- A critical challenge addressed was differentiating between a vessel leaving the ROI and a vessel remaining moored but temporarily silent (due to signal loss or transceiver shutdown). To solve this, the authors implemented a "constrained transit area" approach. They defined successive transit zones extending from the bay's entrance, applying dynamic time thresholds ( $t_{thr}$ ) based on vessel speed and distance from the receiver. A vessel was only classified as absent if its gap in data exceeded a calculated threshold dependent on its speed and the specific transit zone, effectively distinguishing true departures from temporary signal gaps.
High-Resolution Metric Computation:
- Vessel activity was reconstructed at a spatial resolution of approximately 30 meters ( $1'' \times 1''$ ).
- Metrics included vessel counts, transit rates, speed, and course direction.
- Berth and anchorage identification utilized a density-thresholding approach on the gridded vessel density map, an alternative to traditional cluster-based methods. This allowed for the identification of 161 active berths across seven major port areas.
Receiver Localization via Radio Shadowing:
- The study leveraged the "radio shadow" phenomenon, where tall urban buildings occlude AIS signals between vessels and receivers. By fitting geodesic lines to the edges of these signal shadows in the AIS message distribution, the authors inferred the precise locations of passive receiver stations.
- Using a spherical statistical approach (Kent distribution and Hotelling $t^2$ statistic), they calculated the focal points of these shadow patterns to estimate receiver positions with high precision, even for stations not officially reported by the network.

Key Results

Vessel Activity: The analysis revealed an average of 381 vessels operating simultaneously in Tokyo Bay, with 35 actively moving and 346 stationary (berthing, anchoring, or fishing). The daily transit rate (entries/exits) was estimated at $293 \pm 22$ vessels.
Trend Analysis: The data confirms an accelerating long-term trend toward fewer but larger commercial vessels. In 2024, at least 20% of transiting vessels exceeded 10,000 Gross Tonnage (GT), a significant increase from 11% in 2006. The average gross tonnage of transiting vessels was approximately 11,860 GT.
Berth Identification: The study identified 193 berth and anchorage locations, with 161 within 200 meters of the shore. The busiest berths were identified for tugboats, container ships, and tankers, with specific locations in the Yokohama-Kawasaki port area mapped with high fidelity.
Receiver Localization: The "Tokyo receiver" (the dominant source of data) was localized to within a 68% containment range of 1.5 km by 390 m, aligning closely with a known station (AISHub 3752) despite the network not listing it. A second "Yokohama receiver" was also inferred, though with lower precision.

Significance and Claims
The paper claims that open-access AIS data, when processed with rigorous statistical methods to account for coverage gaps and signal occlusion, can map port dynamics with a precision comparable to official statistics.

Data Utility: The study demonstrates that open-access data, including contributions from amateur radio operators, can provide insights into vessel traffic and berth activity that rival proprietary or government sources. This enables high-resolution, near real-time analysis of port performance and economic trends.
Privacy Implications: A significant finding is that in dense urban environments, the radio shadow patterns in AIS data can reveal the precise locations of passive receivers. The authors argue that this capability challenges the anonymity of operators who wish to share data without disclosing their physical location, suggesting that data privacy for terrestrial AIS receivers may be fundamentally compromised in urban settings.
Methodological Contribution: The paper introduces a robust density-thresholding approach for berth identification and a refined method for handling temporary data gaps in vessel status classification, which are applicable to other coastal regions with uneven receiver coverage.

The authors conclude that these findings support strategic and operational planning for shipping industries and port operators while offering a tool for independent monitoring of macroeconomic trends and environmental targets. However, they remain modest regarding the limitations, acknowledging that the study covers only a three-month period and that future work should integrate SAR and AIS-B data to account for "dark" (non-transmitting) vessels.