Cristian-Ioan Vasile, an assistant professor of mechanical engineering and mechanics at Lehigh University, has received a Faculty Early Career Development (CAREER) Award from the U.S. National Science Foundation (NSF). The five-year award will support Vasile’s research on improving the reliability and predictability of machine learning-enabled robotic systems in complex real-world environments.
Vasile, who is also affiliated with Lehigh’s Autonomous and Intelligent Robotics (AIR) Lab, is developing structured methods for evaluating the capabilities of autonomous systems such as drones, robotic assistants, and self-driving vehicles. His research aims to enhance how these agents are deployed in dynamic settings by enabling clearer assessments of their behavior and performance.
According to Vasile, the integration of advanced hardware and machine learning software has introduced new challenges in robot planning and coordination. While such systems can perform complex tasks in controlled environments, their behavior in unpredictable real-world conditions remains difficult to assess. A central issue, he explains, is that many machine learning algorithms operate as opaque systems, making it hard to determine how or why they function as they do.
The research will focus on developing a framework to characterize an agent’s capabilities in terms of motion, manipulation, and perception. The goal is to create interpretable capability profiles that relate a robot’s hardware and software configurations to performance indicators such as task completion time and energy consumption, taking into account variables such as location, time, and environmental context.
The project includes three major components. The first involves building a formal model of capability profiles that link an agent’s performance to specific contexts. The second is the development of planning strategies that reflect a continuum of capabilities rather than binary assumptions. The third focuses on mechanisms for detecting and adapting to failures in task execution, enabling systems to reassign tasks or adjust plans in real time.
Vasile’s work is intended to support the safe and efficient use of autonomous agents in various sectors, including logistics, health care, and service industries. Rather than replacing human labor, he notes, such technologies may be used to address labor shortages and assist in physically demanding or hazardous environments, particularly in regions experiencing demographic shifts.
