Latency Reduction and Packet Synchronization in Low-Resource Devices Connected by DDS Networks in Autonomous UAVs

Area of Science:

• Robotics and Control Systems
• Aerospace Engineering
• Computer Science

Background:

• Real-time flight controllers increasingly rely on general-purpose operating systems.
• Non-deterministic OS introduce critical weaknesses in motion control due to random delays.
• High-speed UAV dynamics require near-deterministic communication for optimal performance.

Purpose of the Study:

• To assess communication latency between companion computers and RTOS open-source flight controllers.
• To compare XRCE-DDS with standard MAVROS-MAVLink communication methods.
• To develop a latency estimation and correction algorithm for improved synchronization.

Main Methods:

• Implemented an XRCE-DDS bridge between low-resource clients and the ROS2 DDS network.
• Measured and compared communication latency statistics.
• Developed an exponential moving average filter-based algorithm for latency offset and clock skew estimation.

Main Results:

• XRCE-DDS demonstrated advantages over MAVROS-MAVLink in communication latency.
• The proposed algorithm effectively estimates latency offset and clock skew.
• The method provides a tool for developers to enhance synchronization.

Conclusions:

• The XRCE-DDS bridge offers a viable solution for reducing communication latency in UAVs.
• The latency estimation algorithm aids in synchronizing time-critical processes.
• This work addresses challenges posed by non-deterministic OS in GNC applications.