Jiawei (Janna) Lin

I am a first-year Ph.D. student in Computer Science and Engineering at University of California, San Diego, where I work on safety and reliability in autonomous systems. I am co-advised by Prof. Sean Gao and Prof. Sylvia Herbert.

Prior to joining UCSD, I graduated with a Bachelor of Science in Computer Science with Honors and Minor in Robotics in the College of Engineering at Cornell University. During my undergraduate studies, I've explored a wide range of topics in autonomous systems control. I was fortunate to work with Prof. Sarah Dean on model-based control for high-altitude balloons, Prof. Hadas Kress-Gazit on continuous execution of high-level collaborative tasks for heterogeneous robot teams, and Prof. Tapomayukh Bhattacharjee on whole-arm physical human-robot interaction (CoRL 2025).

jal214 at ucsd dot edu  /  GitHub  /  Google Scholar  /  LinkedIn

News

Research

I'm interested in the safety and reliability in autonomous systems with learning-based or optimization-based methods.

High-Altitude Balloon Station-Keeping with First Order Model Predictive Control

We revisit the assumption about the efficacy of model-based control for station-keeping by developing First-Order Model Predictive Control (FOMPC). By implementing the wind and balloon dynamics as differentiable functions in JAX, we enable gradient-based trajectory optimization for online planning. FOMPC outperforms a state-of-the-art RL policy, achieving a 24% improvement in time-within-radius (TWR) without requiring offline training, though at the cost of greater online computation per control step.

PrioriTouch: Adapting to User Contact Preferences for Whole-Arm Physical Human-Robot Interaction

We introduce PrioriTouch, a framework integrating a novel contextual bandit algorithm with hierarchical operational space control to learn user contact preferences and translate them into low-level pose and force control policies.

Continuous Execution of High-Level Collaborative Tasks for Heterogeneous Robot Teams

We propose a control synthesis framework for a heterogeneous multi-robot system to satisfy collaborative tasks, where actions may take varying duration of time to complete.

Other Projects

Car Dynamics Correction with Random Fourier Ridge Regression and Fast Optimal Control via DLRA Scheme of Hamilton-Jacobi-Bellman Equation: a Comparative Study

Using the EU long-term dataset, we develop and compare three approaches: raw INS integration, an Extended Kalman Filter with Random Fourier Features (EKF-RFF) for state estimation, and a tensor- train-based Hamilton-Jacobi-Bellman (HJB) solver with Dynamical Low-Rank Approximation (DLRA) for optimal control synthesis.

Robot-assisted Dressing for Closed Geometry Clothing

We propose a grasp planning framework, created using pushing actions, that bunches up cloth in a manner similar to how humans handle cloth during dressing.

Miscellaneous

Website last updated Feb 2026 Design and source code from Jon Barron's website