Marine autonomy and USV electronics
A sealed hull has no convection path and a mast has four radios competing for the same sky. Marine autonomy is a thermal and RF-coexistence problem long before it is a compute problem.
Evidence
What we can substantiate
Everything below is work we have actually done, described at a level that respects the client’s confidentiality. If you need detail, ask under NDA.
- Fanless thermal in a sealed box, sized properly No convection means the enclosure wall is the only heat path, and the internal ambient sits well above the outside air. We build the resistance chain from junction to case to internal air to wall to sea air, size the conduction path to the wall deliberately, and derate the processor against the internal ambient rather than the datasheet's.
- Mast coexistence — AIS, GNSS, telemetry and comms together AIS transmits at 161 MHz within meters of a GNSS antenna and a cellular or satellite modem. Harmonics and intermodulation products land in each other's bands. We plan the frequency inventory first, then separation, filtering and duty-cycle interlocks — because two radios that each pass alone will not necessarily coexist.
- Sensor bandwidth budget before the compute is chosen An autonomy payload is a data-rate problem. We inventory every sensor by interface, rate, latency requirement and duty cycle, add headroom for what the customer will bolt on in two years, and only then pick silicon. TOPS figures say nothing about whether four camera streams fit through the ISP.
- Connectors and corrosion, treated as design work Salt and condensation attack anything unsealed, and dissimilar metals in a wet bilge corrode galvanically. Sealed circulars with 360° shield termination, bonding planned rather than incidental, drain paths and breather vents, and conformal coating specified with the rework path in mind.
- Condensation is the failure nobody designs for A sealed enclosure that cools overnight condenses whatever moisture it was assembled with. A breather with a hydrophobic membrane, assembly humidity controlled, and creepage assessed for a wet surface rather than a dry one.
- Graceful behavior when the vessel loses power Dying-gasp reporting, clean filesystem shutdown, and enough hold-up energy to complete the longest realistic write — sized against the actual flash program time, not a rule of thumb.
Common questions
Straight answers
What actually fails first in a marine enclosure?
Thermal, then connectors, then condensation. A sealed box has no convection, so a processor rated for 25 °C ambient may be sitting in 60 °C internal air and throttling or failing early. Connectors follow — salt and condensation attack anything not properly sealed and shield-terminated. Then the one people miss entirely: an enclosure sealed on a humid day condenses inside every night for the rest of its life. All three are architecture decisions, not layout ones.
How do you stop AIS from blinding the GNSS?
You plan for it rather than discover it. AIS transmits with real power at 161 MHz a short distance from a GNSS antenna on the same mast; harmonics and mixing products can land near L1, and a receiver that is fine in isolation loses satellites the moment the transponder keys. Separation and antenna placement first, then filtering, then — if the geometry genuinely will not allow it — a duty-cycle interlock so the two are never active simultaneously.
Do I need marine-specific certification?
It depends entirely on the vessel class and what the equipment does. Navigation-critical equipment on a crewed vessel falls under a much heavier regime than an autonomy payload on an experimental platform. Establish which applies before the architecture is fixed, because the answer changes isolation, redundancy and documentation requirements — and retrofitting those is expensive.
Got a board to design — or one that won’t boot?
You talk to the engineer who would do the work. Reply within one business day, and we’ll sign your NDA before you go into detail.