Side Projects

Here is a collection of small academic projects, primarily for coursework at Stanford University. Most are not published elsewhere, although some eventually ended up in referred journals or conference proceedings.

DR-iLEQG: Distributionally-Robust Optimal Control of Nonlinear Dynamical Systems for Safety-Critical Applications

• Project for AA 222: Engineering Design Optimization, Stanford University, Spring 2019-20
  • Instructor: Prof. Mykel Kochenderfer
• Proposed novel distributionally-robust optimal controller for nonlinear dynamical systems. For complete version, see this paper.
• Received AA 222 Spring 2020 Best Paper Award.

Human Trajectory Prediction in Socially Interacting Crowds

• Project for CS 230: Deep Learning, Stanford University, Fall 2018-19
  • Instructor: Prof. Andrew Ng
• Presented deep convolutional neural network (CNN) approach to predicting mutually-interacting human trajectories in dense crowd. [Link to Project Report]
• Received CS 230 Fall 2018 Outstanding Poster Award. [Link to Poster]

Information-Theoretic Belief-Space Planning for Gaussian Systems with Nonlinear Observations

• Project for AA 203: Introduction to Optimal Control and Dynamic Optimization, Stanford University, Spring 2017-18
  • Instructor: Prof. Marco Pavone
• Proposed novel belief-space planning algorithm based on model predictive path integral control (MPPI). Showed that existing MPPI framework could be applied to Gaussian belief systems with nonlinear observations if underlying system dynamics are linear. [Link to Project Report]

Dynamic Occupancy Grid Filtering for Multi-target Tracking and Goal Location Inference

• Project for AA 273: State Estimation and Filtering for Aerospace Systems, Stanford University, Spring 2016-17
  • Instructor: Prof. Mac Schwager
• Proposed framework for multi-target tracking that simultaneously estimates unknown goal locations for dynamic agents, e.g. pedestrians. [Link to Project Report]

Decentralized Multi-target Search and Coverage Using Hyperparameter Consensus

• Project for AA 277: Multi-Robot Control, Communication, and Sensing, Stanford University, Winter 2015-16
  • Instructor: Prof. Mac Schwager
• Presented linear-consensus-based algorithm for estimating unknown positions of multiple targets using team of sensing agents. [Link to Project Report]