AnotherPeak presents the smart electric cruise-ship project. Onboard AI and decentralized SQL on EdgeLake, no cloud dependency, across the fleet operating on the French-Swiss Alpine lakes.
Lionel, technical director at AnotherPeak in the French Alps, walks through the "Eagallion" project: a fleet of roughly 13 electric tourist boats spread across three Alpine lakes, recently upgraded from diesel to hybrid-electric and now running EdgeLake as the data platform for battery management and pilot assistance.
The customer's brief was incremental. Start cheap and simple, but build for the future. The two immediate goals are battery lifetime (the batteries are the boat's main asset now) and pilot assistance during cruises. EdgeLake sits between the boats and the maintenance back-office, with most of the heavy logic running onboard.
Each boat's industrial layer is a mix of Modbus and CANbus speaking to the battery-management system, the generator, and the protective relays. AnotherPeak does not connect EdgeLake directly to those binary buses. A small data-logger device sits in front of each subsystem, converts industrial protocol to HTTP/JSON, and feeds EdgeLake. That keeps the edge layer protocol-agnostic and lets the team adopt new vendors without rewiring the data fabric.
The 4G constraint in the middle of an Alpine lake is what forced the architecture choice. There is not enough bandwidth to stream raw telemetry continuously, so the AI inference and the state-of-charge calculations run locally on the boat, and EdgeLake pushes only the decision outputs (route, weather-adjusted state-of-charge, alerts) to the pilot and the maintenance team on shore. Geolocation traces are tracked for regulatory reasons and also fed back into the data-science loop, where weather context lets the model calibrate observed battery behavior against expected behavior for each cruise.