Robotics and Autonomous Systems

Using autonomous surface, underwater, and air vehicles our team is developing a system capable of individual and coordinated actions relative to feature maps, and adaptive sampling. Part of our work is to incorporate new acoustic communication, novel sensing techniques, and adaptive navigation systems to improve the performance and safety of our autonomous operations capable of individual and coordinated actions using feature maps and adaptive sampling.


Deployment of a Kayak in Singapore coast.

The Surface Craft for Oceanographic and Undersea Testing (SCOUT) were developed by engineers at MIT’s Department of Ocean Engineering in order to facilitate surface and undersea oceanographic monitoring. The SCOUTs use ocean kayaks equipped with computers, electronics and mechanical propulsion mechanisms. In addition, the vehicles can be equipped with various third-party and custom oceanographic, vision, navigational, and acoustic sensors. The kayaks are largely used as a test bed for robotic algorithms including but not limited to cooperative multi-vehicle operation and intelligent obstacle avoidance.


AUV in operation.

The Iver-2 family of vehicles from Oceanserver are a modern workhorse AUV designed for coastal data collection. The system is designed to give a single operator the ability to run several vehicles, including launch and retrieval. The vehicles are modified at CENSAM to be equipped with oceanographic chemical sensors, acoustic transducers for communication and localization and ADCP. The vehicles use their built in sensors such as pressure sensors, altitude meters, surface GPS and DVL for navigation.

Adaptive Path Planning and Inspection


A jetty and the model constructed from the scan data.

Adaptive path planning and inspection is motivated by the various incidents such as collapsed bridges, oil spills, and marine structure failures which cause human loss and damage the natural ecosystem. The basic idea is to prevent failures thorough structural inspection. In order to do that fundamental algorithms and robotic technology for autonomous surface reconstruction of marine structures has been developed. In particular, we are developing algorithms for robust robot motion planning, robotics technology for autonomous navigation and scanning of marine structures and software technology to construct 3D models of marine structures.


Biomimetic Locomotion


3D sketch of a biomimetic AUV.

Pervasive sensing in harsh environments requires mobile vehicles with high maneuverability, mechanical robustness, high propulsive efficiency, and stealth. Biological locomotion techniques is ideal as they carry out the potential of enabling all these features. Understanding the governing physics that enable biological propulsive performance can help us adapt such techniques to man-made vehicles and improve their performance. Mission oriented prototypes are Larger scale, Communication & Control (Wifi modem, GPS, compass, IMU), Sensors (Scanning sonar, temperature, salinity) and Power supply by Li-ion cells.


Environmental Model


Thermal image of an industrial outfall.

Environmental models are linked to tidal, hydrodynamic, chemical and biological factors of the atmosphere. We try to gather these information using our robotic fleet. Data acquired by this system enable us to study the environmental parameters of the Singapore coastal area. Our group uses a mixed physics-based and learning-based approach to model various oceanographic phenomenon such as harmful algae blooms.