Roboat

At MIT Senseable City Lab, the Roboat project focused on developing a fleet of modular, self-driving boats for the canals of Amsterdam. The work spanned robotics, embedded systems, and urban sensing, aiming to adapt autonomous driving technologies to a highly unpredictable aquatic environment. I focused on the concept design, naval architecture, and physical prototyping, translating open-ended research goals into testable, multi-directional surface vehicles capable of passenger transport and environmental monitoring.

The prototyping pipeline involved scaling the system through one-meter, two-meter, and full-scale four-meter iterations. I built the initial models using 3D-printed hulls and RC jets, eventually integrating gyroscopes and a custom drive-by-wire system to automatically adjust thrust for stabilization. The vehicles utilized a symmetrical, single-hull design with interlocking connectors, allowing multiple units to dock and form floating platforms. To monitor the fleet, I developed a remote-control architecture and a full-stack telemetry visualization platform that streamed real-time sensor data during deployments.

Operating on water introduced severe constraints, as unpredictable currents and lidar refraction made navigation and object detection highly complex. Early testing revealed that traditional catamaran layouts limited internal cargo capacity. To resolve this, I collaborated with a naval engineer to run hydrodynamic simulations, validating that a single-hull design could achieve the required stability while maximizing modular space. Testing involved debugging the drive-by-wire system to compensate for lateral drift and refining the sensor arrays to differentiate between submerged debris and swimmers under dynamic environmental conditions.

The system was deployed as full-scale prototypes in Amsterdam for passenger transport and waste collection. The research and hardware development contributed to multiple IEEE publications, including "Roboat II: A novel autonomous surface vessel for urban environments" and "Design, modeling, and nonlinear model predictive tracking control of a novel autonomous surface vehicle". The physical artifacts and system architecture were also exhibited at the Museum of the Future in Dubai as part of the Tomorrow Today exhibition.