Building a Real-Time Testing Platform for Unmanned Ground Vehicles with UDP Bridge

Area of Science:

• Robotics and Autonomous Systems
• Computer Simulation and Graphics
• Vehicle Dynamics and Control

Background:

• Autonomous driving faces significant perception and control challenges, necessitating robust testing methodologies.
• Realistic simulations are crucial for testing autonomous systems in safety-critical and inaccessible scenarios.
• Existing simulators often overlook non-urban environments, which present unique challenges for unmanned ground vehicles (UGVs).

Purpose of the Study:

• To develop a flexible, real-time testing platform for UGVs in indoor and off-road settings.
• To address the limitations of current simulators that predominantly focus on urban driving scenarios.
• To facilitate efficient testing and algorithm validation for autonomous vehicle development.

Main Methods:

• Developed an original simulator integrated with the Robotic Operating System (ROS).
• Implemented a novel communication solution, the UDP Bridge, for real-time data exchange.
• Generated real-life interchangeable data for seamless simulation-to-reality transfer.

Main Results:

• The integrated platform enables real-time execution of autonomy algorithms directly within ROS.
• The UDP Bridge demonstrated up to 9.5 times faster communication compared to existing solutions like ROS#.
• Achieved successful experimental validation of autonomy algorithms using the developed platform.

Conclusions:

• The proposed platform offers a flexible and efficient solution for testing UGVs in diverse, challenging environments.
• The UDP Bridge significantly enhances communication speed, enabling real-time performance for complex autonomous systems.
• This work contributes a valuable tool for advancing research and development in autonomous driving, particularly for non-urban applications.