Satellite Startup Tracks Ships' Radars and Radios From Space
Virginia-based startup HawkEye 360 has come up with a unique idea to bolster maritime domain awareness. The firm uses a constellation of small satellites to detect and locate the source of commercial radio frequency emissions – everything from VHF push-to-talk radios to maritime radar transmissions, AIS beacons, satellite mobile comms and more.
HawkEye recently partnered with Amazon’s ML Solutions Lab to incorporate machine learning algorithms into their analytics. Using AWS's Amazon SageMaker Autopilot, they generated AI models to be used for an automated maritime vessel risk assessment process.
Why is this revolutionary? Because it has the potential to uncover hidden patterns and relationships among vessel features that previous ML algorithms failed to do. Potential vessel behaviours of interest include illegal fishing, human trafficking, ship-to-ship transfers, sanctions-busting, GPS jamming and smuggling.
When bad actors turn off their AIS signal to hide their vessel’s position, HawkEye can still spot the traces they leave by tracking their radar and their VHF calls. “RF signals can provide valuable insight into commercial vessel activity across the globe, even when some seek to hide their location,” said HawkEye 360 vice president of products Tim Pavlick. “With these machine learning-backed capabilities, we will empower customers to cut through an ocean full of noise to obtain more timely and critical insights from maritime RF data to improve mission outcomes and prevent illegal and illicit activities.”
The ability to use ML algorithms to counter illicit drug smuggling would be a first for the industry, even though the same approach has been implemented on land. The reason is that there are so many variables and complexities involved in the process for seagoing targets. With Amazon’s help, HawkEye 360 says that it has overcome these challenges.
“By combining HawkEye’s data and deep domain expertise with Amazon SageMaker Autopilot, HawkEye 360 is able to halve the time for machine learning model development and deployment. That frees up time for data scientists to focus on creating new and innovative solutions to the world’s problems,” said Amazon's senior manager for machine learning solutions, Sri Elaprolu.