Optical Flow Module Positioning Control Algorithm

Optical flow positioning control algorithms use downward-facing cameras and distance sensors to estimate velocity and maintain stable UAV position, often enhanced with IMU data and robust control methods like PID or Sliding Mode Control.

Overview of Optical Flow Positioning

Optical flow modules estimate the motion of a UAV relative to the ground by analyzing sequential images from a downward-facing camera. These sensors are often paired with a distance sensor, such as LiDAR, to provide altitude information, enabling accurate velocity estimation even in GNSS-denied environments like indoors or underground . The optical flow output measures angular rotations of the ground image, which are then converted into translational velocities along the UAV axes .

Sensor Integration

Modern optical flow modules, such as the PX4Flow, ARK Flow MR, or Holybro H-Flow, integrate a camera, distance sensor, and a 6-axis IMU. The IMU provides angular velocity and acceleration data, which is fused with optical flow measurements to improve position and velocity estimation . Some advanced setups use a gimbal-mounted optical flow sensor to stabilize the camera independently of UAV attitude, reducing errors caused by pitch and roll movements .

Control Algorithms

PID Control

For simpler implementations, optical flow data is integrated to estimate position, and a PID controller adjusts rotor speeds to maintain hover and counteract drift. This method is commonly used in modules like the LiteWing Flight Positioning Module, where the UAV autonomously maintains its position using optical flow and dead reckoning .

Sliding Mode Control

For more robust performance, especially in dynamic or GPS-denied environments, Sliding Mode Control (SMC) can be applied. SMC stabilizes both the UAV attitude and the optical flow module independently, ensuring the sensor remains level while the UAV maneuvers. The gimbal is controlled via rotor torques to maintain horizontal alignment, improving velocity and position estimation accuracy .

Guidance Algorithms

Advanced guidance algorithms process optical flow data with additional parameters like relative depth to reduce noise and disturbances. These algorithms allow UAVs to navigate corridors or confined spaces autonomously, avoiding collisions while maintaining precise positioning .

Practical Implementation

1. Sensor Setup: Mount a downward-facing optical flow camera and a distance sensor. Optionally, use a gimbal for stabilization.
2. Data Fusion: Combine optical flow measurements with IMU data to estimate velocity and position.
3. Control Loop: Implement PID or Sliding Mode Control to adjust rotor speeds and gimbal angles, maintaining stable hover and position.
4. Autonomous Position Hold: Integrate dead reckoning or guidance algorithms to maintain position in GNSS-denied environments .

Applications

• Indoor UAV navigation and hovering
• Aerial photography requiring stable position
• Autonomous inspection in confined or GPS-denied areas
• Research platforms for testing advanced control algorithms By combining optical flow sensors, distance measurements, IMU data, and robust control algorithms, UAVs can achieve precise positioning and stable flight even in challenging environments.