ESE PhD Candidate Farhad Savvas defended his thesis on Hierarchical Planning for Safe and Reactive Robot Autonomy in Dynamic, Human-centric Environments Tuesday, June 30th. Advised by Professor Nikolai Matni and Professor Nadia Figueroa, Savvas’ research focuses on ensuring safety for robotics technology through a bi-level planning network.
“Current robotics technology often struggles with safety guarantees and computational tractability when deployed in unpredictable, human-centric environments like homes, the service industry, and busy parking lots,” Savvas explains. “My research bridges this gap using a bi-level planning framework across both manipulation and navigation: we combine high-level task planning algorithms that generate reactive, adaptable behaviors with low-level motion planning modules that mathematically guarantee safe and stable physical movements.”
Throughout his research, Savvas experienced his share of challenges. “Robotics is an inherently interdisciplinary field, demanding a seamless blend of expertise from computer science, mechanical engineering, and electrical engineering, thus making it difficult to translate algorithms from scratch to diverse physical robots such as arms and autonomous vehicles,” Savvas explains. “I overcame these hurdles by leaning heavily on collaboration. Working closely with my mechanical engineering peers and industry partners in Honda Research Institute (HRI), we were able to translate theoretical concepts from the ideation phase all the way to robust, real-world hardware implementations.”
PhD advisors, Prof. Nikolai Matni and Prof. Nadia Figueroa also played a significant role in ameliorating the above challenges for Savvas. So did ESE faculty mentors Professor Santosh S. Venkatesh and Professor Manfred Morari “They were all instrumental in my growth as a researcher and student,” Savvas emphasizes. “I would also like to extend a heartfelt thank you to all my lab peers, co-authors, and collaborators who made this journey so rewarding.”
Savvas describes his five-year academic journey as one marked by remarkable professional and personal growth.
“A standout milestone was during my first year, passing the qualification exam in probability theory. It was an intense, abstract math course, and getting through it was a truly collaborative effort with my peers in Electrical and Systems Engineering (ESE),” Savvas says, noting this milestone, “initially felt completely out of reach.”
In addition, the two teaching practicums Savvas completed with Professor Venkatesh and Professor Morari underscored the significance of his degree milestones. “Collaborating with and learning from such inspirational faculty members was a defining part of my graduate journey.”
As for future plans, Savvas will join Lila Sciences as a Scientist I, “focusing on Robotics, Dexterous Manipulation, and Robot Learning. My long-term vision is to continue pushing the boundaries of intelligent control and machine learning, making advanced robotics technology safer, more capable, and genuinely accessible to everyday people.”
When Savvas isn’t working, he enjoys biking on the Schuylkill River Trail, playing tennis, and walking. “Philadelphia is such a vibrant, walkable city and I really enjoy just wandering through different neighborhoods and exploring what the local food scene it has to offer.”
Learn more about Savvas’ work here