Skip to content

NON_GPS_NAVIGATION.md

Latest commit

 

History

History
218 lines (158 loc) · 5.85 KB

NON_GPS_NAVIGATION.md

File metadata and controls

218 lines (158 loc) · 5.85 KB

Non-GPS Navigation Documentation

Setup and configuration guide for GPS-free navigation using visual-inertial systems with ArduPilot.

Original Documentation: Google Docs

Hardware Requirements

  • Flight controller running ArduPilot (Pixhawk, Cube, etc.)
  • Down-facing camera (mono or stereo)
  • IMU sensor (integrated in flight controller or external)
  • Companion computer (for running ROS2 nodes)
  • Calibrated camera-IMU system (see CAMERA_IMU_CALIBRATION.md)

Software Requirements

  • ArduPilot firmware (Copter, Plane, or Rover)
  • ROS2 (Humble or later)
  • MAVROS2 or MAVLink
  • NGPS, UKF, and VIO packages from this repository

Wiring & Connection

Camera Connection

  • Connect camera to companion computer via USB, Ethernet, or MIPI CSI
  • Ensure camera provides timestamped images
  • Configure camera frame rate (typically 10-30 Hz)

IMU Connection

  • Use flight controller's onboard IMU (recommended)
  • Or connect external IMU via I2C/SPI to companion computer
  • Ensure IMU data is accessible via ROS2 topics

Companion Computer to Flight Controller

  • Connect via serial (USB, UART) or Ethernet
  • Configure MAVLink connection
  • Set appropriate baud rate (921600 or higher recommended)

ArduPilot Configuration

Enable External Navigation

  1. Set EK3_ENABLE = 1 (enable EKF3)
  2. Set EK3_SRC1_POSXY = 3 (external position source)
  3. Set EK3_SRC1_VELXY = 3 (external velocity source)
  4. Set EK3_SRC1_POSZ = 1 (barometric altimeter for altitude)
  5. Set EK3_SRC1_VELZ = 1 (barometric altimeter for vertical velocity)

Configure Position Source

Set EK3_SRC1_POSXY = 3 to use external position from MAVLink VISION_POSITION_ESTIMATE message.

Configure Velocity Source

Set EK3_SRC1_VELXY = 3 to use external velocity from MAVLink VISION_SPEED_ESTIMATE message.

Other Important Parameters

EK3_GLITCH_RAD = 5.0          # Position glitch protection radius (meters)
EK3_REJECT_GPS = 0            # Don't reject GPS (if available)
AHRS_EKF_TYPE = 3           # Use EKF3

MAVROS Configuration

Basic Setup

  1. Install MAVROS2:

    sudo apt install ros-humble-mavros ros-humble-mavros-extras

  2. Configure connection:

    ros2 run mavros mavros_node \
  --ros-args \
  -p fcu_url:=/dev/ttyUSB0:921600 \
  -p system_id:=1 \
  -p component_id:=1

Topic Mapping

MAVROS publishes/subscribes to:

  • /mavros/odometry/out: Fused odometry output (to ArduPilot)
  • /mavros/imu/data: IMU data (from ArduPilot, optional)
  • /mavros/vision_pose/pose: Vision position estimate (to ArduPilot)

Enabling and Configuring IMU & Camera Data Streams

Methods to Enable High-Resolution IMU

Method 1: Via MAVLink Streams

In ArduPilot, configure high-rate IMU streams:

SR0_ADSB = 0
SR0_EXT_STAT = 2
SR0_EXTRA1 = 10
SR0_EXTRA2 = 10
SR0_EXTRA3 = 2
SR0_PARAMS = 10
SR0_POSITION = 3
SR0_RAW_SENS = 2
SR0_RC_CHAN = 0

Method 2: Via Companion Computer

Read IMU directly from flight controller via I2C/SPI if supported.

Full Setup Steps

  1. Start MAVROS:

    ros2 launch mavros apm.launch.py fcu_url:=/dev/ttyUSB0:921600

  2. Start VIO system:

    ros2 launch ap_vips vio.launch.py

  3. Start NGPS localization:

    ros2 launch ap_ngps_ros2 ngps_localization.launch.py \
  reference_image_path:=/path/to/satellite_image.tif

  4. Start UKF fusion:

    ros2 launch ap_ukf estimator.launch.py

  5. Verify topics are publishing:

    ros2 topic list
ros2 topic echo /ukf/odometry

Calibration Procedure

Steps

  1. Calibrate Camera-IMU system (see CAMERA_IMU_CALIBRATION.md)

    • Camera intrinsics
    • IMU noise parameters
    • Camera-IMU extrinsics
    • Time synchronization

  2. Configure ArduPilot parameters:

    • Set EKF3 as primary estimator
    • Configure external navigation sources
    • Set appropriate noise parameters

  3. Test on ground:

    • Power on system
    • Verify all ROS2 nodes are running
    • Check that odometry is being published
    • Verify ArduPilot is receiving external navigation

  4. Initial flight test:

    • Start in manual/stabilize mode
    • Verify position estimates are reasonable
    • Check for drift and accuracy
    • Monitor EKF health in Mission Planner/QGroundControl

  5. Tune parameters:

    • Adjust EKF noise parameters if needed
    • Tune UKF fusion weights
    • Optimize NGPS matching parameters

Notes

  • Always test in a safe, open area first
  • Keep GPS enabled initially for comparison and fallback
  • Monitor EKF health and innovation values
  • External navigation requires good camera visibility and lighting
  • System performance depends on altitude (higher is better for NGPS)
  • Ensure adequate computational resources on companion computer
  • Regular calibration recommended (every few months or after hardware changes)

Troubleshooting

ArduPilot Not Receiving Position

  • Check MAVROS connection: ros2 topic echo /mavros/state
  • Verify odometry topic is publishing: ros2 topic echo /ukf/odometry
  • Check EKF3 source parameters in ArduPilot

Poor Position Accuracy

  • Verify camera-IMU calibration
  • Check camera focus and image quality
  • Ensure adequate lighting conditions
  • Verify reference image quality and georeferencing

High Drift

  • Check IMU noise parameters
  • Verify time synchronization
  • Ensure sufficient visual features in camera view
  • Check for camera vibration or motion blur

System Crashes

  • Monitor companion computer CPU/memory usage
  • Reduce camera frame rate if needed
  • Optimize feature matching parameters
  • Check for overheating

Additional Resources