14 DISC
Maritime Security Analytics Software (MSAS) with Artificial Intelligence and Machine Learning for Coastal Surveillance Network of Indian Coast Guard
The following areas have been selected for integration of AI/ML in CSN software:-
3.1 Classification of Radar Data. Presently, Radar Data provides kinematics and positional information of the vessel. It does not classify the type of vessel, activities and interaction between vessels. Hence, AI/ML software is required for classification of type, activities and interactions between vessels at sea.
3.2 Automatic Detection, Classification by EO Sensor and Association with Radar/AIS Track. Presently, the EO sensors are manual and required to be operated by watchkeepers. There is no automatic capturing of image of Radar/AIS track, classification of type & identity w.r.t. library, tagging to track and alerting on mismatch if any. Hence, AI/ML software is required for automatic capturing without operator intervention, tagging to track, building image library for classification of type, activities, interactions between vessels and alert on mismatch/interaction.
3.3 Multilingual ASR & NLP Solution for VHF Radio communication.
VHF Radio is installed on all Radar Stations and software has inbuilt feature to record the conversation. Presently, there is no software to covert the analogue voice to digital and text form. Hence, multilingual ASR & NLP Solution is required to convert all Indian languages(especially used in coastal states) voice to digital form in English which can be searched by key word to generate intelligence form open broadcast.
3.4 Predictive Analysis for AIS tracks. Presently, AIS Data provides
kinematics, positional information and some static information of the vessel as transmitted on AIS. However, the software is not able to classify authenticity of static information transmitted on AIS and activities of the vessels. Hence, AI/ML software is required for anomaly detection in static information and vessel activities based on track.
3.5 Retrieval Augment Generation (RAG) Application for PANS Data.
Presently, PANS data is being rendered in PDF, Word, XLS format on mail to ICG. The PANS data has vital information but could not be stored is structured database and fused to AIS database due to format. Hence, AI based RAG Application for converting Unstructured data (PANS and other sources) to structured database and fusing the information to AIS data for generating risk intelligence. Further, software should be compatible for integration of e-PANS from National Logistics Portal (Marine) of Indian Port Association as and when available.
3.6 AI Based MDA Assistant for Helping Watchkeeper to Investigate Suspicious Vessels. Presently, watchkeeper investigate the vessel over radio for suspicious movement. However, operator may miss certain activities of the vessel. Hence, AI based MDA Assistant may be designed to find out factors of suspicion for investigation of the vessel and initiation of AI based investigation if required.
