Using GNSS data and AIS, WAVES will enable ship tracking, estimate ships’ time of arrival, and model port situations allowing for more efficient harbor management leading to faster transport and less port congestion.
Port congestion is a major problem in maritime transportation. Port congestion leads to massive economic losses wherein both businesses and consumers are affected due to delays in the delivery of goods. Businesses also lose a portion of their income through additional fees and penalties caused by the delays. Meanwhile, the supply of products may remain limited affecting consumers (Economic Impact of the Port Congestion, n.d.). The National Customs Brokers and Forwarders Association of America has even named port congestion as an “economic threat” (Port Congestion: An Economic Threat, n.d.) To illustrate, in 2014, the Philippines lost 2.5 billion (42 million euros) per day due to port congestion. This also caused the country's agricultural goods also decrease in competitiveness due to additional logistical costs (Ordonez, 2016)
WAVES aims to address the problem. Under Challenge #3: Marine Applications, WAVES will be used for positioning applications with the goal of making seaport operations more efficient.
WAVES caters to the needs of three types of users: (1) shipping lines (2) port authorities and (3) the public.
Using the application, shipping lines and port authorities will be able to:
Track vessels’ locations;
Predict a ship’s route and time of arrival; and
Model the port situation during a vessel’s time of arrival; and
Provide recommendations based on an AI-assisted decision making model.
Meanwhile, as the region is composed of archipelagic countries, marine transportation remains a major mode of transporting people and goods. WAVES will make it easier for commuters and businesses their estimate their ship’s time of arrival and current location while in transit.
WAVES’ can ease the problem of port congestion because it will predict the route and time of arrival of ships and recommend docking and embarkation schedules ensuring ports will not be congested at any given time. Because port authorities can see the port situation during their time of arrival, they can advise shipping lines regarding their route and velocity. This will be realized by utilizing GNSS and AIS and employing a machine learning algorithm.
Robust and high precision model in the prediction of vessel routes and ETA and the high-level AI-assisted decision making as the core of the WAVES application sets it apart from the current players in the Market and also from existing Government tools. WAVES will set a precedent to all other innovations that aim to improve the processes in the Maritime Industry.
The technical back-end of WAVES works by integrating three main components namely, GNSS, machine learning, and port-side data. Our application connects directly with Galileo and AIS to determine a shipping vessel’s location. WAVES takes advantage of Galileo’s interoperability with other GNSS constellations to allow better triangulation of vessel position in real-time. Once a vessel has been triangulated, ship information such as longitude, latitude, speed, heading, and destination are transmitted to Galileo. These data are then authenticated by Galileo to prevent falsified information before sending it to our application. The transmitted information is then processed by our second component, a machine learning (ML) model. Further details of the succeeding components can be found in Step 7. As shown in Figure 1, WAVES is at the center of all these procedures helping both port and vessels simultaneously by providing real-time information on ETA, and recommendations on changes in speed and route.
Figure 1. WAVES' transfer of data.
Determining vessel location at sea is an important task for the timely transport of goods, improvement of navigation, bringing down safety and security risks, and, ultimately, reducing costs of anchoring. Our application aims to address these issues through the integration of Galileo. With our Galileo-enabled application, real-time updates on vessel location and velocity are used to predict its route and estimated time of arrival (TOA) to port and improve scheduling with other vessels. These features of WAVES are highly dependent on accurate and true data.
As noted in a GNSS market report in 2015 , positioning accuracy and data integrity are two of the most crucial factors in navigation. Errors in estimating vessel position could cause hazards in navigation and cause unnecessary costs. Galileo’s capabilities aim to address these concerns by providing a high-accuracy service, and authentication of signal and data. Accurate positioning data is possible due to Galileo’s modern signals, which prevents multipath, and interoperability with other GNSS constellations.
Transportation plays a key role in the movement of goods and people. 80% of traded goods around the world are being carried by ships (UNECTAD, 2021). According to the International Chamber of Shipping, around 11 billion tons of products are transported by ships annually. In 2019, the total value of the world shipping trade reached more than USD 14 trillion (International Chamber of Shipping, n.d.). Thus, seaport management plays a vital role to ensure a continuous and seamless global trading system.
As an archipelagic region, marine transportation is one of the most common ways of transporting people and goods in ASEAN. WAVES aims to focus on the maritime shipping transport sector and improve seaport management. There are two main thrusts of WAVES: 1) provide a real-time and accurate time of travel for ships from its port of origin to its destination, and; 2) serve as a monitoring and alert system for seaports as a decision-making policy tool to avoid port congestion.
It focuses on three main customers: (1) shipping lines; (2) seaport authorities; and (3) businesses and people who use maritime transport. All of these groups have the need to get real-time information regarding a ship’s route and time of arrival. Moreover, a mobile application and web service will be developed to cater to the different target customers.
The application shall employ a freemium policy model wherein its basic application of providing estimated time travel and geolocation for ships shall be free to all users. Meanwhile, WAVES’ annual premium tier on monitoring and alert systems are intended for its primary users (shipping lines and seaports) in the maritime and logistics sector. The premium tier of WAVES is estimated to cost around $154.00 annually that includes acquisition of data, the cost for creating the model through machine learning and artificial intelligence, and maintenance cost for the monitoring and alert systems. Specifically in the ASEAN region, WAVES is estimated to have an annual total addressable market (TAM) worth of $14,188.13 for servicing 47 major ports (ASEAN, 2002) alone. WAVES’ TAM is still expected to significantly increase once utilized by various shipping lines, logistics and related companies, non-major seaports, and even the commuting public for their navigation and tracking of ships and cargoes and other business use.
WAVES builds on the seafaring history of our Southeast Asian ancestors and taps into the growing utilization of GNSS and other existing technologies in marine transportation that aims to enhance port productivity. Seaports and shipping lines will be provided of an estimated time of departure and arrival (ETDs/ETAs) of ships navigating from port A to port B and vice versa. WAVES can also track the vessels’ location through the use of GNSS and AIS data. Likewise, given the ETDs and ETAs of shipping lines navigating on the seas, seaports of origin and destination will simultaneously receive the same information and provide decision-making tools for managing the seaports to avert possible economic losses due to port congestion.
With better seaport management, port productivity will be enhanced and ease of mobility of goods shipped within the ASEAN region is expected to be achieved. This would further create a multiplier effect on various sectors of the economy of each country in the region. Businesses will be able to better manage their supplies based on the tracking location and time of arrival in the port of destination of the vessels.
As seen in Figure 1, the central idea of WAVES is to employ a machine learning algorithm to predict the route and TOA of ships, and use this algorithm in support of higher decision-making algorithms. This WAVES model will consider other factors such as port clearance speed and volume of shops at bay to recommend docking and embarkation schedules, and assist port management in changing vessel route and velocity. The ML model is first trained on a publicly available dataset which makes use of Automatic Identification System (AIS) for marine traffic. From this initial dataset, the model will be able to predict an estimated route and TOA based on a ship’s current position, heading, and speed. The predicted TOA will then be used to construct a timetable of arriving ships per port. This timetable takes into account not only satellite and vessel information, but also port-side information such as the docking capacity, customs clearance speed, discharging speed, and queueing/holding facilities. ML-assisted vessel routing and speed modelling are also provided with port-specific timetables to properly guide port management for the arrival of ships. This trained ML model will be used as the initial algorithm for WAVES and will be constantly updated based on port-side stored data.
The last component of our application is integration with port-side data. The predictions, schedules, and recommended changes by our application are transmitted to the destination port. Data received by port management are then used to prepare for the arrival of ships, and communicate changes. Changes in vessel route and velocity are communicated from port to ship through the WAVES platform. These changes include new routes, heading, and speed. All of these components aim to reduce port congestion, manage marine traffic, and, ultimately, reduce costs of anchoring.
Machine Learning Model (1st Level)
Figure 2. Ship route and ETA prediction model
Thousands of observations from both publicly and privately available datasets on ship location, speed, direction, and marine weather data will be run(?) on a computer program to build a route and ETA model for ships. The route model for the first program will be based on the computations made from raw data of the ship's location. A sample mathematical model can be built for the two outputs:
ETA = f(latitude, longitude, velocityt, velocityt-1,weather) +
Route(latitudet+1,longitudet+1) = f(latitudet, longitudet, velocityt, velocityt-1,weather) +
The second program will process historical route visualization of routes taken by ships from one port to another. The program will detect patterns on the routes not addressed by the calculations. This program will be integrated with the first program to arrive at a more robust prediction model.
Figure 3. Decision Making Model from Galileo Information to Timetables and Recommendations
Once the Route and ETA Model has been built, a simulated real time location data will be processed, with other factorial data such as port clearance speed, weather forecast, vessel types on the harbor etc., in order to create sets of decision-making models for port management. This will enable WAVES to create recommendations to both the ships and the management all the while using the application as a form of communication. Once users, both from the ports and the ships, reach a decision, WAVES will recalculate and offer voyage solutions to the ships.
WAVES will mainly use real-time location data from AIS-dependent databases. To ensure consistent performance of our solution, our application will also utilize data received from Satellite Based Augmentation Systems available in the East and Southeast Asia in order to correct errors in geopositioning of the vessels.
Since AIS data comes from an independent source, the GNSS-enabled WAVES application can use its own location to check for positioning errors and improve accuracy of the ships location in the application interface.
The WAVES application will rely heavily on already available methods for prediction, especially machine learning programs and softwares in order to build a robust and accurate model. Artificial Intelligence might also be employed in building the computational model for decision-making.
A hardware requirement for the WAVES app to work in ships are telecommunications devices that can enable internet connectivity. This is important to ensure real time update of ships' locations and provide better accuracy in terms of route and ETA projections. These will play as important factors in the decision-making capacity of WAVES.
The figures below show an early demonstration of our product in web application and mobile application form. Figure 4 shows an interface of vessels arriving at a particular port. Additionally, a timetable of arriving ships is shown on the left side with vessel ID, time of arrival, and next ship in queue. For a more detailed perspective, Figure 5 shows an AI-recommendation column which guides port-management in dealing with port-traffic. This column lists details of changes in vessel speed and routing. It depends on management's decision whether these changes will be set in place or whether to disregard these recommendations. For approved changes, waves will automatically send routing and vessel speed information changes to corresponding ships. As mentioned previously, WAVES aim to properly manage port congestion and reduce costs of anchoring. With an automated scheduling and recommendation system, port management can focus on optimizing other processes such as customs clearance procedures and discharging. For vessels, they experience a seamless coordination with ports regarding changes in velocity and routes.
Figure 4. WAVES Web Application
Figure 5. WAVES Mobile Application User Interface
The WAVES project team is composed of five individuals from different backgrounds. This ensures that everyone brings something to the table and can contribute to different aspects of the project development. Collectively, the team has significant experiences in quantitative and qualitative research, project management, international cooperation, business development, GIS, design, machine learning, and statistics. The team's educational background and professional profile ensures the quality of the outcome.
Leo Jaminola graduated with a degree in Political Science and has worked in research and project management for the past three years. He is currently pursuing a master's degree in Demography at the same university.
Frank Cally Tabuco is currently a Master’s of Science in Computer Science student at University of the Philippines Diliman. His main interest of research is on medical image analysis and bioacoustics using deep learning.
Mariano Niño V T. Oliva is a data analyst undergraduate with special interests in machine learning and I.T. applications in game and app development.
Ralph Chester D. Retamal is a Graphic Designer and GIS Specialist focusing on transportation, urban planning, and sustainability.
Monica Paula Lavares is an economist and attained her master’s degree in Economics at Ateneo de Manila University. Her interests include research on public health, transportation, and digital technologies.