Real-Time Optical Communication Using Event-Based Vision with Moving Transmitters
A new system uses event cameras to send text between moving robots with 95% accuracy, 7x faster than prior methods.
A research team has developed a novel optical communication system designed to solve the radio frequency (RF) congestion and jamming problems plaguing multi-robot teams. Their system, detailed in a paper submitted to IROS 2026, abandons traditional frame-based cameras in favor of event-based vision sensors. These specialized cameras offer microsecond temporal resolution and exceptional performance in high dynamic range lighting, making them uniquely sensitive to the rapid light changes from a moving optical transmitter. This allows the system to maintain a reliable communication link even during fast relative motion between robots.
The core of their breakthrough is a custom tracking and decoding pipeline built around a Geometry-Aware Unscented Kalman Filter (GA-UKF). This algorithm is the key to the system's performance, enabling it to achieve over 95% decoding accuracy for text transmission while the transmitter is in motion. Critically, the team's method processes this data 7 times faster than the previous state-of-the-art technique, all while maintaining equivalent tracking accuracy at high transmitting frequencies of 1 kHz and above. This combination of speed and precision is what makes real-time, robust communication between fast-moving drones or ground robots a practical reality.
This work represents a significant step toward freeing robot swarms from the limitations of RF spectrum. By using light as a communication medium, teams of robots could operate with greater autonomy and coordination in environments where RF is unreliable, contested, or simply too crowded. The demonstrated high accuracy during motion opens the door for complex command-and-control and data-sharing applications in dynamic, real-world scenarios.
- Uses event cameras for microsecond resolution & high dynamic range, overcoming motion blur and lighting issues of standard cameras.
- Achieves >95% text decoding accuracy during motion and processes data 7x faster than prior state-of-the-art methods.
- Implements a Geometry-Aware Unscented Kalman Filter (GA-UKF) for robust tracking of transmitters at frequencies ≥ 1 kHz.
Why It Matters
Enables reliable, jam-resistant communication for drone swarms and robot teams operating in dynamic or contested environments.