Autonomous systems are increasingly reshaping industries such as transportation, logistics, healthcare, and construction. From self-driving vehicles to delivery robots and drones in construction, these intelligent machines are transforming how we live and work. But for all their promise, deploying robots in real-world, dynamic environments remains one of the greatest challenges in robotics.
At the forefront of solving this problem is Cristian-Ioan Vasile, assistant professor of mechanical engineering and mechanics at Lehigh University and faculty member in the Autonomous and Intelligent Robotics (AIR) Lab. Recently, Vasile earned a prestigious National Science Foundation (NSF) CAREER Award to advance research that makes autonomous robots more predictable, reliable, and effective in complex environments.
Why predictability matters in robotics
With advances in artificial intelligence, machine learning, and robot perception systems, robots are increasingly capable of sensing, reasoning, and acting within their surroundings. However, performance in controlled training environments does not always translate seamlessly into the real world.
“Many machine learning algorithms for robotics are opaque,” explains Vasile. “They perform well under training conditions, but it is often unclear how they will behave in unfamiliar or dynamic scenarios.”
This lack of transparency limits wider adoption of autonomous systems in construction, logistics, and healthcare. For instance, while hospitals use robots for simple delivery tasks, these machines operate with restricted autonomy due to uncertainty about their performance in sensitive situations.
Building capability profiles for autonomous agents
Vasile’s research focuses on developing structured methods for mapping and modelling a robot’s capabilities. These capability profiles capture how robots perform under varying contexts—such as navigating crowded spaces, operating in low-light environments, or adapting to sudden changes.
The framework will link hardware, software, and environmental context to measurable outcomes like energy consumption, task completion time, and reliability. By moving away from binary assumptions of “capable” or “not capable,” Vasile aims to create a performance spectrum that better reflects real-world complexity.
For example:
- In a construction site, can a robot safely navigate narrow corridors filled with moving equipment?
- In a grocery store, how effectively can it stock shelves or manoeuvre through crowded aisles?
- In healthcare, can it deliver medication accurately even in busy hospital corridors?
Such insights will help engineers, planners, and operators better allocate robotic systems in large-scale, team-based deployments.
Smarter planning, faster recovery
The project’s second phase involves designing advanced planning methods that consider contextual performance rather than relying on simplistic “yes/no” assessments. This means accounting for whether a robot performs differently at noon versus night, in a quiet corridor versus a crowded warehouse.
A third critical element is failure detection and recovery. If a robot is misassigned or underperforms, the system should detect the mismatch, adapt its role, or reassign the task in real time. This adaptability is essential for scaling up the safe use of autonomous robots in construction, logistics, and beyond.
Implications for industry and the workforce
While discussions about robotics often raise concerns about job displacement, Vasile emphasizes that the goal is not to replace humans but to augment human capabilities.
“Many of the tasks that robots could do are in areas facing worker shortages or involve dangerous, strenuous, or repetitive work,” he notes. In sectors such as construction—where labour shortages and safety concerns are significant—robots could play a pivotal role in performing hazardous or physically demanding tasks, allowing human workers to focus on higher-value, creative, and supervisory roles.
Toward scalable, reliable robotic systems
The NSF CAREER-funded research represents a significant step toward scalable, predictable, and context-aware robotic systems. By improving how robots plan, adapt, and collaborate in real-world environments, Vasile’s work lays the foundation for widespread adoption of autonomous systems in construction, healthcare, logistics, and transportation.
As industries continue to embrace digital construction technologies and AI-driven automation, smarter planning for autonomous systems will be essential to unlocking their full potential.
Read next: A decade of digital transformation in construction – Build in Digital
Are you a building professional? Sign up for a FREE MEMBERSHIP to upload news stories, post job vacancies, and connect with colleagues on our secure social feed.
