Skip to content

Roadmap

We have the core hardware in hand — ESP32-CAM modules for the submerged vision unit, Heltec Wireless Trackers for the smart buoy and base station. The software stack is open-source tooling: TensorFlow Lite Micro for on-device inference, Arduino/ESP-IDF for firmware, LoRa for communications.

Active

  • Train and validate the species classification model on the ESP32-CAM
  • Build and test the servo door latch mechanism
  • Integrate the submerged unit with the Heltec Tracker buoy over a wired tether
  • Bench-test the full bidirectional command protocol
  • Develop base station firmware with fleet management

Goal: Demonstrate end-to-end operation on the bench — catch event triggers classification, buoy relays result over LoRa, base station displays status and sends commands.

Planned

  • Waterproof enclosure design for submerged unit
  • Waterproof enclosure design for smart buoy
  • Shallow-water field trials (dock-side, controlled environment)
  • Open-water field trials with commercial crabbers
  • Validate classification accuracy in real-world conditions
  • Test LoRa range in maritime environment
  • Battery life and solar charging validation

Goal: Prove the system works in real ocean conditions with real crabs.

Planned

  • Design ballast release mechanism
  • Integrate surface command with mechanical actuation
  • Buoyancy testing and ascent rate characterization
  • Safety validation (no hazard to boats, divers, or marine life during ascent)

Goal: Eliminate the buoy line entirely. Pot surfaces on command from the base station.

Planned

  • Wave energy harvester prototype (linear electromagnetic generator)
  • Hybrid solar + wave power management
  • Extended deployment testing (weeks without maintenance)
  • Ultra-low-power firmware optimization

Goal: Self-sustaining buoy that operates indefinitely without battery swaps.

Future

  • Collapsible hydrodynamic fairing design
  • Pneumatic launch rail prototype
  • Fairing-to-trap unfolding mechanism at depth
  • Integration with ropeless recovery for full no-rope workflow

Goal: A crabber never handles rope at all — deploy by launch, recover by command. Dramatically faster deployment cycles, smaller deck footprint.

Planned

  • Phone GPS → NOAA CO-OPS tide table lookup for deployment location
  • Flood/ebb tide alerts (productive window opening and closing)
  • Catch rate correlation with tidal phase (per-pot, per-location)
  • “Pot full” threshold alerts based on keeper count
  • Idle pot detection (no catch events over configurable window)
  • Safe-to-soak confirmation for leaving pots across tidal cycles

Goal: The system tells you when to pull — not your gut. Tidal awareness means you work the productive window and leave pots safely across cycles because GPS + ropeless = never lose a pot.

Future

  • Cloud ingestion pipeline for fleet telemetry
  • Web dashboard for fleet-wide analytics
  • Aggregated catch data API for fisheries managers
  • Catch-per-tidal-cycle analysis and deployment optimization
  • Seasonal pattern analysis across locations and conditions
  • Regulatory compliance reporting

Goal: Move from per-boat operations to fishery-wide intelligence. Catch data correlated with tidal phase, location, season, and conditions creates optimization insights that used to take a lifetime on the water to develop — and gives fisheries managers real-time stock data instead of delayed survey estimates.

Future — Research

The logical endpoint of SmartPot: the cage is the vehicle. Instead of carrying pots to the fishing ground by boat, each pot propels itself from the dock, navigates to a GPS waypoint, submerges, fishes, and either surfaces for pickup or returns home under its own power.

  • Propulsion system selection: electric thruster, wave-powered, or hybrid
  • Steering and rudder mechanism (or differential thrust with dual motors)
  • GPS waypoint navigation firmware (depart dock → transit → arrive at fishing ground)
  • Collision avoidance (AIS receiver, ultrasonic proximity, or camera-based)
  • COLREGS compliance: navigation lights, radar reflector, AIS transponder for autonomous vessel rules
  • Controlled-descent ballast system (flood chamber to sink, blow to surface)
  • Transition from surface navigation mode to bottom-sitting trap mode
  • Reuse existing SmartPot systems at depth: vision, classification, door control, telemetry

Three tiers, from simplest to most autonomous:

ModeHow it worksComplexity
Surface and holdPot surfaces at fishing ground, operator picks up by skiffLow — just ropeless recovery with GPS
Herd modeOne powered “shepherd” ASV tows a string of surfaced pots back as a raftMedium — one smart vehicle, many dumb floats
Self-returnEach pot navigates itself back to dock after fishingHigh — full round-trip autonomy
PhaseEstimated drawDurationEnergy
Surface transit (1-2 mi)50-100W30-60 min50-100Wh
Descent5W (pump/valve)2-5 min~1Wh
Fishing (standard SmartPot)80mA avgHours to daysExisting solar/battery
Ascent5W (ballast blow)2-5 min~1Wh
Return transit50-100W30-60 min50-100Wh

A 14.8V 10Ah LiPo pack (148Wh, ~400g) covers a full round-trip with margin. Combined with existing solar charging, multi-day deployments are feasible between transits.

The current stack provides a surprising amount of the foundation:

  • GPS navigation — Heltec Tracker with L1/L5 GNSS (surface positioning)
  • LoRa command/control — bidirectional, encrypted (waypoint updates, recall commands)
  • Ballast release — Phase 3 ropeless recovery mechanism (reused for dive/surface)
  • Solar power — smart buoy solar harvesting (recharges between trips)
  • Vision system — could augment surface obstacle detection during transit
  • Propulsion hardware (thruster, motor controller, waterproof seals for moving parts)
  • Steering mechanism and navigation firmware
  • Dynamic waterproofing (seals that survive repeated surface/submerge cycles)
  • Regulatory: USCG autonomous vessel rules, COLREGS navigation light requirements
  • Insurance and liability framework for unmanned surface vessels

A static SmartPot targets ~$160/unit. An autonomous drone pot would likely cost $800-2,000/unit — significantly more hardware and complexity. But the economics may still close:

  • Eliminates fuel costs for deployment runs ($50-200/trip for a commercial boat)
  • Eliminates boat time — the single most expensive resource in a crabbing operation
  • Enables fishing grounds beyond casual day-trip range from shore
  • A fleet of drone pots deploying from a dock pier is a fundamentally different business model: no boat required

Goal: A crabber walks to the end of a dock, taps a phone screen, and a fleet of pots deploys itself to the fishing ground, fishes the tide, and comes home full. No boat, no rope, no fuel, no empty pulls.