Portable side scan sonar systems—often integrated into autonomous underwater vehicles (AUVs) and unmanned underwater vehicles (UUVs)—are now routinely deployed to survey the ocean and seafloor for objects of interest, including mines, black boxes, shipping containers, pipelines, human bodies, hull defects, and more. However, surveys are often constrained by the need to physically transport the data back to the surface for analysis, which then involves laborious data inspection by experts.
Charles River’s Solution
Charles River Analytics has partnered with Teledyne Gavia to develop an app, AutoTRap Onboard™, for Teledyne’s industry-leading Gavia line of AUVs. AutoTRap Onboard enables AUVs and other sonar systems to automatically detect and classify objects from side scan sonar data during their missions, in real time. AutoTRap Onboard applies recent advances in deep learning object detection and representation to detect a wide range of objects.
Real Time is Critical
AutoTRap Onboard’s ability to run in real time is novel and significant—other apps for object detection run only during post-mission analysis. Real-time detection enables AUVs to reach a next-generation level of autonomy, where guidance and navigation software can make decisions directly from incoming data. AutoTRap Onboard empowers AUVs to not just collect data but act on it in real time; for example, rerouting towards an object for closer inspection or returning to the surface once a high-priority target is found. These capabilities can reduce costs and widen survey areas for AUVs. Outside of AUV applications, AutoTRap Onboard can support operators of towed sonar systems, who also require the ability to act immediately on detections.
New Directions for Charles River
AutoTRap Onboard is the latest technology to transition out of the lab at Charles River Analytics, a company regularly recognized for its successes in solving near-intractable problems for its government clients.
“Converting our cutting-edge research into robust solutions that positively impact real-world users is central to our mission,” said Richard Wronski, Vice President of the company’s Sensing, Perception, and Applied Robotics Division.
“More and more companies are coming to us,” said Elaine B. Coleman, Vice President of Commercialization. “It’s not just that we do autonomy—we can help them figure out what level of autonomy will work for their customers.”
Coleman and Wronski are confident in AutoTRap Onboard’s potential for commercial success based on the company’s extensive experience developing AI-based tools for the underwater domain.
“From a technology perspective, it’s a great fit for us. Machine learning and AI are squarely in our wheelhouse. And collaborating with premier industry partners like Teledyne Marine Systems to bring this technology to the market is a home run for us, and for the operators, who will reap immediate benefits from using the system,” Wronski said.