MSL Home
Research
Staff
Publications
Facilities
Visit MSL
Contact MSL
 
Research


  • Agile underwater vehicles for monitoring, reconnaissance and sampling
  • Novel acoustic detection and classification systems
  • Infrared and video camera for vessel tracking and threat assessment
  • Cognition-centric integration of observations and models

The above noted core research activities are described below:

Agile underwater vehicles for monitoring, reconnaissance and sampling

Through a partnership with iRobot, we are deploying underwater vehicles for complex missions.  We are working with Rangers andTransphibians (Fig. 1) to implement communications innovations and payload enhancements.  These modifications allow the vehicles to interact with the MSL passive acoustics detection and classification system.  We optimize vehicle paths by utilizing a navigation and mission planning algorithm with input from a high-resolution forecast model of the estuary and coastal currents.  This allows the vehicle to conserve battery life in high current regimes by navigating to places with weaker currents and to overall swim a smarter path through fluctuating currents (Fig. 2). 

Figure 1: Transphibian Underwater Vehicle

Figure 2: Underwater Currents

Novel acoustic detection and classification systems

MSL has designed, built and tested a portable passive acoustics hydrophone detection system coupled with a classification algorithm (Fig. 3).  The system has been fielded in multiple coastal and estuarine settings and has successfully detected divers, swimmers and underwater vehicles at significant distances.

Figure 3: Navy boat with Stevens acoustics mooring on dock at Newport

Infrared and video camera for vessel tracking and threat assessment

A probability threat matrix constructed using the tools of decision science has been integrated into the real-time video feed.  We also utilize infrared (IR) cameras to visualize threats that are difficult to resolve at night or in fog.  The kayakers in the field of view of the IR image (Fig. 5) were undetectable by other means.

 

Figure 4: Gaming environment for assessment of Hostile Intent

Figure 5: IR cameras (left) and Kayakers in background (right)  

Cognition-centric integration of observations and models

MSL is participating in a layered space and land-based system of threat detection across multiple scales covering the open ocean, coastal region and local harbor (Fig. 6). Our vision is for these sensors to form a cognitive network. For example, cognitive sensors are capable of autonomously representing their domain (e.g. parsing AIS data to determine ferry routes and schedules) and sharing distributed information to the local sensor field.  Collectively the sensors can determine if conditions are adverse for a certain type of modality (e.g., too foggy for video feeds) and switch to a more appropriate modality. Cognitive sensors respond cooperatively without human intervention (e.g. sensed anomaly in the water triggers an underwater vehicle deployment).  Cognitive systems learn to interpret using locally sensed data and respond according to the goals of the enterprise.

More on this Stevens-wide research area: Cognition-centric networks

 

Figure 6: Layered Approach

The individual assets described above produce an integrated picture of the maritime domain (Figure 7 and 8).

 

Figure 7: MSL Data Integration

Figure 8: Vessel Tracking and Acoustic Retrievals