Together with the Fraunhofer Center for Maritime Logistics and Services (CML), Fintraffic’s Vessel Traffic Services (VTS) has launched a research project to further develop VHF maritime radio traffic using automatic speech recognition. In the future, a voice recognition system using artificial intelligence (AI) will convert and provide real-time voice radio traffic in text form to Fintraffic’s vessel traffic centers. This will support the correct interpretation of messages, especially in extraordinary situations.
“Fintraffic’s Vessel Traffic Services will be on standby around the clock for the VHF radio system that covers the whole coast of Finland. The VHF talk radio is one of the central tools at our Vessel Traffic Centres. The recent marine VHF traffic-related development work enables connecting machine learning artificial intelligence alongside traditional radio traffic which we are now going to research further in Finland,” says Olli Soininen, program manager at Fintraffic’s Vessel Traffic Services.
”In the research project which is starting now, we will investigate the suitability of utilizing artificial intelligence and automation in Finland’s VTS. We will also build a proto model on a national level and use it to test automatic speech recognition as a support for normal radio traffic, among other things,” continues Mika Nyrhilä, project manager at Fintraffic’s Vessel Traffic Services.
Fintraffic’s Vessel Traffic Services aims to test this prototype as early as 2024. In addition to the VHF voice recognition model, the development project will provide a basis for creating one-way VHF voice announcements with a speech recognition software that uses real-time vessel traffic situational pictures. The automatic transmissions created will be tested for the first time in navigational warnings of the Finnish coastal radio station Turku Radio in 2024.
Artificial intelligence for correct transcription even under difficult conditions
The research project combines the Fraunhofer Center for Maritime Logistics and Services’ (CML) previous experience in speech recognition with Fintraffic’s expertise in the operation and specifics of maritime VHF communication.
”There are multiple things typical to maritime communication, which make understanding and receiving messages more difficult. Oftentimes, there is engine noise drowning out speech, and crews are made up of individuals from various countries with their own accents and dialects. These are all issues that seed to be solved in order to improve maritime communication,” says Maximilian Reimann, research associate at Fraunhofer CML.
In maritime transport, the VHF voice radio system has been used for decades for communication between ships and ports. With the VHF radio, ships can establish a direct voice connection with a coastal station or another ship, as well as contact emergency services. The importance of the system to vessel safety is such that it is mandatory on all major ocean-going vessels. However, the quality of VHF communications at sea is affected by several factors. Among other things, the distance between sender and receiver, antenna heights and their condition, and varying meteorological conditions can affect the clarity of the message.
Source: Fraunhofer CML