Tech United — RoboCup Middle Size League

Overview

As a member of Tech United's research team, I contributed to the development and integration of real-time control systems for wheeled soccer robots competing in the RoboCup Middle Size League (MSL). This is a multidisciplinary university team working on autonomous mobile robotics, computer vision, and multi-agent coordination.

My Contribution

Building on my BSc thesis work, I continued the implementation and integration of the shared human-robot control framework onto the physical TURTLE robots. This involved:

Adapting the control algorithms from simulation to real hardware with actual sensor noise, communication delays, and mechanical constraints
Working within the team's real-time software architecture (RTDB-based) to ensure reliable performance at the control loop rate
Collaborating with team members across perception, planning, and hardware subsystems

Technical Challenges

Real-Time Communication

The shared control commands needed to arrive within the 30 ms control cycle. We used RTDB channels with priority-based scheduling to ensure deterministic delivery:

# RTDB channel write with timestamp
def publish_command(rtdb, robot_id: int, velocity: tuple[float, float, float]):
    """Publish velocity command to a specific robot via RTDB."""
    channel = f"ROBOT_{robot_id}_VEL_CMD"
    rtdb.put(channel, {
        "vx": velocity[0],
        "vy": velocity[1],
        "vtheta": velocity[2],
    })

Sensor Fusion on Hardware

On the physical robots, we dealt with noisy wheel encoders and vision-based positioning. The control system needed to be robust to measurement outliers:

\hat{v}_k = \alpha \cdot v_{\text{encoder},k} + (1 - \alpha) \cdot v_{\text{vision},k}

where $\alpha$ is a confidence weight based on signal quality metrics.

Competition Experience

Contributing to a competitive robotics team provided invaluable experience in:

Working under tight deadlines with hardware-in-the-loop constraints
Debugging real-time systems where failures happen on the field
Interdisciplinary collaboration across software, electrical, and mechanical engineering

Technologies Used

Python, C++, RTDB, real-time control systems, multi-agent robotics