Donipolo Ghimire

Currently:

• Ph.D. candidate in Mechanical and Aerospace Engineering at UC Irvine (since 2020)
• Working under the guidance of Prof. Solmaz Kia

Research: My research focuses on robotics and multi-agent systems, specifically:

• Developing algorithms for coverage and monitoring with collaborative multi-robot systems
• Integrating multiple disciplines: -Computer vision, Optimization, Control theory, Probabilistic machine learning, Variational inference, Graph theory
• Designing advanced planning and deployment strategies for robotic cooperation

Previously:

• Summer 2024: Research Intern at US DEVCOM Army Research Lab
  • Focus: Active source seeking and exploration algorithms for search-and-rescue mission.
• Fall 2022: Visiting Researcher at NASA Jet Propulsion Laboratory
  • Focus: Perception and motion planning for robotic exploration.
• Master in Mechanical and Aerospace Engineering from UC Irvine (2022)
• BSc in Mechanical Engineering from Howard University (2019)

Research Vision: Looking ahead, I am excited to explore the role of abstractions/representation in achieving efficient and robust autonomy in robotics. I believe that the “right” representations are paramount in enabling robots to understand complex environments and execute sophisticated tasks efficiently. I am particularly interested in two approaches. The first is a hierarchical architecture that involves a low-level control loop for basic movement execution (for example, moving from point A to B) and a higher-level system that operates at a slower duty cycle to break down missions into sub-goal. The second is an end-to-end learning architecture that directly maps sensor inputs to motor commands (“pixels to volts”). The hierarchical framework provides the benefit of interpretability and data efficiency, but the end-to-end framework provides more flexibility that learns complex mapping directly from data and leads to emergent behaviors that are difficult to engineer explicitly in a hierarchical system. Ultimately, I aim to and want to encourage the robotics community to develop a hybrid system that integrates structured reasoning with learned flexibility of end-to-end learning, bridging low-level perception and high-level task semantics for resilient and intelligent autonomy. [See more]

I enjoy maintaining an active lifestyle through:

• Hiking
• Climbing
• Listening to podcasts