cv

Princeton University

• PhD, Mechanical and Aerospace Engineering
• Francis Robins Upton Fellowship
• Phillips Second-Year Fellowship
• Harari Post-generals Fellowship

Duke University

• BSE, Mechanical Engineering
• BS, Physics
• Certificate, Innovation and Entrepreneurship
• Magna Cum Laude; Graduation with Distinction

Princeton Intelligent Robot Motion Lab

• Derived an information theoretic fundamental upper bound on robot performance, given the robot’s task and sensing capabilities. Generalized and tightened this bound with f-divergence. Showed that the fundamental bound holds for robot systems of increasing complexity and intelligence.
• Proved an upper limit for language-instructed autonomy. Showed that the ambiguity in language itself would lead to imperfect behavior, regardless of scaling of the large language models leveraged. Estimate the upper limit on realistic home robot scenarios.
• Established end-to-end safety assurances for robot navigation via rigorous uncertainty quantification on learned perception modules.
• Hardware demonstration with the Unitree Go1 quadruped robot.
• Using uncertainty quantification, guide active perception and exploration while maintaining safety.

Duke Pratt Fellows

• Develop theories and simulations on control and optimization of lattice modular multilevel converters with serial and parallel connectivity.
• Obtain most efficient control algorithms for lattice converters under requirements including converter size, input/output terminals, and output voltage/current.

Duke Neutrino & Cosmology Group

• Qork on SNEWS (Supernova Early Warning System), produce sky maps with predicted supernova location and uncertainty intervals.
• Analyze neutrino events detected at the Super-Kamiokande Detector.

DukeEngage in Detroit

• Analyzed TechTown programs via interviews with program directors and clients.
• Provided suggestions for improving each program as well as recommendations for potential collaborations across programs.
• Served as peer mentor for 12 student entrepreneurs for Launch Detroit, a summer entrepreneurship camp for technology-oriented college students.

Duke Motorsports Team

• Engaged other students to build and assemble wheels for the car competing in 2020 and 2021.
• Designed, modified, and performed analysis on the wheels for the car competing in 2021.

Fundamental Limits for Sensor-Based Robot Control

• Our goal is to develop theory and algorithms for establishing fundamental limits on performance for a given task imposed by a robot's sensors. In order to achieve this, we define a quantity that captures the amount of task-relevant information provided by a sensor. Using a novel version of the generalized Fano inequality from information theory, we demonstrate that this quantity provides an upper bound on the highest achievable expected reward for one-step decision making tasks. We then extend this bound to multi-step problems via a dynamic programming approach. We present algorithms for numerically computing the resulting bounds, and demonstrate our approach on three examples, (i) the lava problem from the literature on partially observable Markov decision processes, (ii) an example with continuous state and observation spaces corresponding to a robot catching a freely-falling object, and (iii) obstacle avoidance using a depth sensor with non-Gaussian noise. We demonstrate the ability of our approach to establish strong limits on achievable performance for these problems by comparing our upper bounds with achievable lower bounds (computed by synthesizing or learning concrete control policies).

Perceive with Confidence | End-to-End Safety Assurances for Robot Navigation

• Rapid advances in perception have enabled large pre-trained models to be used out of the box for processing high-dimensional, noisy, and partial observations of the world into rich geometric representations (e.g., occupancy predictions). However, safe integration of these models onto robots remains challenging due to a lack of reliable performance in unfamiliar environments. In this work, we present a framework for rigorously quantifying the uncertainty of pre-trained perception models for occupancy prediction in order to provide end-to-end statistical safety assurances for navigation. We build on techniques from conformal prediction for producing a calibrated perception system that lightly processes the outputs of a pre-trained model while ensuring generalization to novel environments and robustness to distribution shifts in states when perceptual outputs are used in conjunction with a planner. The calibrated system can be used in combination with any safe planner to provide an end-to-end statistical assurance on safety in a new environment with a user-specified threshold 1-epsilon. We evaluate the resulting approach — which we refer to as Perceive with Confidence (PwC) - with experiments in simulation and on hardware where a quadruped robot navigates through indoor environments containing objects unseen during training or calibration. These experiments validate the safety assurances provided by PwC and demonstrate significant improvements in empirical safety rates compared to baselines.

• Harari Post-generals Fellowship (2024)
• Phillips Fellowship (2023)
• Francis Robins Upton Fellowship (2022)

• Dean's List with Distinction (2018, 2019, 2021)
• Engineering Honor Societies: Tau Beta Pi, Pi Tau Sigma
• Graduation with Distinction (2022)