VAIL focuses on aerial and aquatic drones.
The Vehicle Autonomy and Intelligence Lab at the Luddy School of Informatics, Computing, and Engineering has received an Amazon Web Services Machine Learning Research Award for the lab’s research to enable drones to make decisions in real-time using streams of sensing data received from onboard sensors such as Light Detection and Ranging (LiDAR), sonar, and cameras.
VAIL’s proposal will develop open-source perception and decision-making libraries which can benefit the machine learning and robotic communities at large. The award, which is worth $70,000 in AWS credits, will allow VAIL researchers to access and leverage Amazon’s state-of-the-art cloud computing tools and services for thousands of hours, boosting their work on both machine learning and autonomous systems.
“Machine learning algorithms are increasingly being utilized by autonomous robots for building various data-driven models,” said Lantao Liu, an assistant professor of intelligent systems engineering at Luddy and the director of VAIL. “The robots typically have weak computational capacity due to their limited dimensions and payloads, yet they require online learning with data processed on the fly. Unfortunately, many algorithms for solving these tasks entail large computational costs which can be very challenging for the robot.”
Utilizing the resources of Amazon’s AWS, such as the Amazona Apache MXNet deep learning software framework and the AWS SageMaker, which is used for training robot perception models, will allow VAIL researchers to focus on innovative models rather than the time-consuming tasks of model training and deployment workflow.
“We are in the process of designing reinforcement learning methods to make the robots behave smarter,” Liu said. “Training RL on these different kinds of robots requires diverse training environments, and SageMaker RL supports easy integration of our customized simulators, which will significantly facilitate our research.”
The AWS Machine Learning Research Awards (MLRA) assists faculty, Ph.D. candidates, and graduate students with research to advance the frontiers of machine learning and its application across a wide range of problems—from finding new therapies for cancer to solving climate change and exploring outer space. MLRA provides eligible researchers and university programs cash awards and AWS promotional credits so that they can do more faster using the most advanced compute, analytics, and machine learning tools available in the cloud.