Real-time Sensing & Control

Enterprises and public sector organizations can benefit from increased use of autonomous vehicles for sensing and data collection, and even to deliver payloads to sites or regions that are difficult for humans to access in a timely manner.  The challenge is to make autonomous platforms flexible, responsive and easy to use, without the need to custom-develop each application.  It is also vital to ensure that the need for human decision making managed at the "mission" level, enabling g the autonomous systems to sense, learn and make decisions locally in order to fulfill the mission.

 Swim ESP for Real-time Sensing and Control uses efficient, real-time, context-aware machine learning to learn locally and automatically – on autonomous vehicles equipped with sensors, on assets at the network edge, and in the cloud.   On assets, Swim finds hidden correlations and patterns in floods of raw data in real-time using self-training edge learning algorithms that run on commodity, low-power devices.  It transforms noisy  data into a low-rate stream of high value learned insights that can be used for fast local decision making, including flight control to track identified objects, collision avoidance, swarm control and many more.  These insights are also available at the network edge, where more sophisticated, mission-specific applications and operators can respond.   

Swim ESP for Autonomous Vehicles (Demo Video)