April 27, 2020
From autonomous cars to robot assistants, intelligent vehicles and robots are quickly becoming a part of our daily lives. But how do we avoid possible collisions and accidents with machines that are programmed to make decisions? And how do we make them better equipped to choose the best trajectory to their final destination?
Efstathios Bakolas, an associate professor in the Department of Aerospace Engineering and Engineering Mechanics, has plans to help answer some of these questions using two grants from the National Science Foundation (NSF) that support research in path planning and trajectory optimization for autonomous robots and vehicles.
The grant awarded for the project “Efficient Algorithms for Safety Guiding Mobile Robots Through Spaces Populated by Humans and Mobile intelligent Machines and Robots” supports research that aims to develop safe and smooth integration between intelligent robots and vehicles used in our daily lives.
Humans interact with a variety of autonomous and semi-autonomous robots in different spaces including dense urban spaces like busy crossroads and factory floors at large industrial facilities. It is important for robots and humans to avoid collisions with each other which could cause physical harm to humans as well as damage to the robots and other surrounding equipment. A key challenge is to develop autonomous robots and vehicles that have the ability to make decisions in real-time and in unpredictable circumstances. Bakolas’ proposed research will create algorithms that allow robots to predict the most likely future motion patterns of surrounding humans and robots in real-time while safely guiding them to their destination without collisions.
The grant awarded for the project “Real-Time Trajectory Generation Algorithms for Uncertain Autonomous Based on Gaussian Processes” supports work that will contribute new theory and algorithms to optimize trajectory generation for autonomous systems.
Self-driving vehicles and robots support the national economy in a number of ways, such as transporting and delivering goods and increasing productivity and efficiency in manufacturing. Similar to predicting the weather or traffic conditions, it is difficult to accurately predict future conditions under which the autonomous system will be operating in real-time. Instead of committing these systems to a single trajectory, Bakolas proposes that creating algorithms and systematic methods to compute a collection of alternative trajectories will help autonomous vehicles and robots arrive to their destinations more quickly and efficiently. His research team will develop model-based and data-driven algorithms that can be executed in real-time based on current conditions.
Bakolas joined the department in 2012. His main research interests include optimal control theory, stochastic control, optimization-based control, differential games and game theory, computational and algorithmic geometry, dynamic programming, nonlinear control, path planning and motion planning for robotic systems and autonomous vehicles, and distributed control and estimation for multi-agent networks. Learn more about his research on his website.
