Modules

Core executable modules and their key components in the JetsonNano ECU system.

Vehicle System (src/vehicle.cpp)

The Vehicle System module maintains the vehicle state model and provides actuation interfaces aligned with VSS (Vehicle Signal Specification).

Key Components:

• VehicleSystem: Central state container with VSS-aligned hierarchy
• Vehicle: Core state representation with sub-systems
• MotionManagement: Steering and powertrain management
• Exterior: External vehicle elements

Responsibilities:

• Maintains current vehicle state
• Subscribes to control commands via Zenoh
• Actuates chassis/body components
• Exposes vehicle state via standardized interfaces

Middleware (src/middleware.cpp)

The Middleware module bridges CAN bus communications with Zenoh, enabling seamless integration between vehicle hardware and software components.

Key Components:

• BatterySensor: Monitors power metrics (voltage, current, power) via I²C
• Signals: Main CAN↔Zenoh bridge for bidirectional communication

Responsibilities:

• Receives current speed from Arduino and publishes it via Zenoh (so the system has up-to-date speed)
• Publishes battery sensor data via Zenoh and CAN
• Subscribes on Zenoh for control commands (e.g., lights) and sends them via CAN
• Relays other control signals (autonomous level, emergency braking alerts, etc.) from Zenoh to CAN or to the Raspberry Pi cluster as needed
• Provides standardized Zenoh interfaces for vehicle communication

Combined Controller (src/combinedControl.cpp)

The Combined Controller handles all vehicle control functions, from manual inputs to autonomous steering and speed.

Key Components:

• XboxController: Reads gamepad inputs, exposes manual speed/steering
• PIDController: Steering control with fixed/predefined desired speed values
• MPCController: Optimizes both steering and desired speed based on trajectory
• SpeedPidController: Throttle control that attempts to reach the desired speed (used with both PID and MPC)

Responsibilities:

• Processes manual inputs from Xbox controller
• Switches control algorithms based on automation mode
• If using PID:
  • PIDController handles only steering based on lane midpoint error
  • Uses fixed/predefined desired speed values
• If using MPC:
  • MPCController optimizes both steering and the desired speed target
  • Determines optimal desired speed based on trajectory/curves
• In both cases:
  • SpeedPidController handles the actual throttle control to reach the desired speed target
• Subscribes to perception outputs (when in autonomous modes)
• Publishes control commands to Middleware

Detection (src/detection.cpp)

The Detection module processes camera inputs to understand the environment, detecting lanes, objects, signs, and lights.

Key Components:

• LaneDetector: Uses a deep learning model to segment and identify lane boundaries from camera frames
• ObjectDetector: Identifies vehicles, pedestrians, signs, and road surfaces
• TrajectoryDefinition: Generates path planning based on lane/object fusion
• TrafficSignClassifier: Classifies detected signs and lights
• AdaptiveCruiseControl: Estimates distances for ACC behavior
• ObstacleAvoidance: Path planning for obstacle avoidance in SAE_4

Responsibilities:

• Captures and processes camera frames
• Runs model inference for lane segmentation
• Calculates lane midpoint error
• Identifies objects and classifies them
• Determines road/non-road segments
• Generates trajectory for autonomous steering
• Classifies traffic signs and lights
• Triggers emergency braking when needed
• Publishes perception outputs to Combined Controller