Technical Specifications

Deep dive into the engineering behind the Odessos autonomous sailboat.

Hardware Architecture

The Odessos prototype is a 1-meter scaled version of the Class 420 sailboat, automated via custom rail and servo systems.

Hull & Structure

Based on the Class 420 design, the hull provides a stable and proven platform for autonomous sailing. It features a fiberglass construction reinforced for mounting custom hardware.

  • Material: Fiberglass / GRP Composite
  • Length: 1.0 meter (scaled prototype)
  • Type: Monohull Planing Dinghy
  • Features: Self-bailing cockpit, reinforced transom
FIG 1.0: HULL_GEO
SCALE: 1:10

Rudder Control

A custom rail-based system driven by a high-torque electric motor and GT2 belt allows precise changes to the rudder angle. This linear motion is converted to rotational force for steering.

  • Actuation: Linear Rail + Stepper Motor
  • Transmission: GT2 Timing Belt
  • Precision: 0.5 degrees
  • Response Time: < 100ms
SYS: STEERING_ACT
STATUS: ACTIVE

Boom & Sail System

An intelligent winch system adjusts the mainsheet to optimize the angle of attack. The boom position is constantly monitored by rotary encoders to ensure maximum lift.

  • Main Sail Area: 0.8 m²
  • Control: High-torque Winch Servo
  • Sensors: Boom Angle Encoder
  • Range: 0° to 90° (Port/Starboard)
MOD: AERO_DYN
WIND_V: VAR

Balance Control

To prevent capsizing in strong winds, an active ballast system moves a lead weight along a transverse rail. A PID controller uses gyroscope data to maintain a level horizon.

  • Counterweight: 2.5 kg Lead Block
  • Rail Length: 400 mm
  • Control Loop: 50Hz PID
  • Max Correction: 15° Heel
SYS: STABILITY
PID: ENGAGED

Sensors

The masthead sensor array provides real-time environmental data. This input is fused with GPS and IMU data to calculate the optimal heading.

  • Wind: Ultrasonic Anemometer
  • Position: RTK GPS Module
  • Orientation: 9-DOF IMU (BNO055)
  • Vision: Wide-angle Camera (YOLO)
DATA: TELEMETRY
LINK: 4G/LORA

Software & Intelligence

Machine Learning (YOLO)

Odessos uses a YOLO (You Only Look Once) model for object detection. It identifies other vessels and obstacles to ensure safe navigation. The model is currently trained on over 0 images, with a dataset of 0 prepared for future refinement.

Simulation & Pathfinding

Our physics-based simulation tests the boat in virtual environments with variable wind, currents, and obstacles.

  • Downwind: Calculates the shortest direct path, avoiding obstacles leeward.
  • Upwind: Performs tacking maneuvers (zigzagging) to move against the wind.
  • Obstacle Avoidance: Dynamically recalculates routes to bypass islands or moving ships.

Physics Core

The movement logic is governed by hydrodynamic and aerodynamic principles.

Max Load Capacity

W_MAX = ρ * V_MAX - M Where ρ is water density (1025 kg/m³)

Resistance Force

F_H = 1/2 * C_H * ρ * V_B² Hydrodynamic resistance calculation

Acceleration

F - R = a * (M + W) Driving force minus resistance equals mass times acceleration