Curriculum Vitae 

Google Scholar

Email: haiminh (at) princeton.edu

My Research Wordcloud:

 

Hello and welcome! I am a third-year Ph.D. candidate in the Department of Electrical and Computer Engineering at Princeton University, advised by Professor Jaime Fisac. My research is centered around learning for control, human-robot interactions (HRI), model predictive control (MPC) and multi-agent systems.

I did my masters in the Department of Electrical and Systems Engineering at the University of Pennsylvania. I was fortunate to work with Professors George Pappas, Manfred Morari and Nikolai Matni in the GRASP Lab. I won the 2020 SEAS Outstanding Research Award.

I received my bachelor degree with honor in Electronic and Information Engineering from ShanghaiTech University, proudly among its first cohort of bachelor students since the university was founded in 2013. During my senior year I studied as a visiting undergraduate student in the Department of EECS at UC Berkeley.

During my undergraduate years, I worked with Prof. Boris Houska at ShanghaiTech University on the topic of tube-based robust MPC via min-max differential inequalities. From 2017 to 2018 I was a research assistant in the Hybrid Systems Lab at UC Berkeley, working on trajectory planning and distributed MPC, advised by Prof. Claire Tomlin.

 

News:

New preprint “Active Uncertainty Reduction for Safe and Efficient Interaction Planning: A Shielding-Aware Dual Control Approach” is out! This work contains results obtained during my internship at Honda Research Institute USA in Summer 2022.

• • The preprint is available on arXiv here.
  • Hardware demonstration video can be found here.

* This page is extremely out of date.

Non-Cooperative Distributed MPC with Iterative Learning

21st IFAC World Congress, 2020

In this project we present a novel framework of distributed learning model predictive control (DLMPC) for multi-agent systems performing iterative tasks. The framework adopts a non-cooperative strategy in that each agent aims at optimizing its own objective. Local state and input trajectories from previous iterations are collected and used to recursively construct a time-varying safe set and terminal cost function. In this way, each subsystem is able to iteratively improve its control performance and ensure feasibility and stability in every iteration. No communication among subsystems is required during online control. Simulation on a benchmark example shows the efficacy of the proposed method.  [Paper]  [Slides]  [Video]

---

Tuning Communication Latency for Distributed MPC

8th IFAC Workshop on Distributed Estimation and Control in Networked Systems (NecSys 2019), Chicago

In this project we develop a distributed MPC scheme for a class of input-coupled linear systems implemented over wireless networks. The proposed approach allows each agent to achieve reduced communication latency by sending less information to their neighbors. Uncertainties incurred by this delayed and incomplete information are handled by a constraint tightening procedure. Simulation examples demonstrate that for distributed systems with chain structure, our tightening method is invariant to the number of agents in the network. Moreover, we show that by a proper tuning of latency, an optimized control performance can be achieved.  [Paper]  [Poster]

---

Min-max Differential Inequalities for Polytopic Tube MPC

2019 American Control Conference, Philadelphia

This project is concerned with robust, tube-based MPC for control systems with bounded time-varying disturbances. In tube MPC, predicted trajectories are replaced by a robust forward invariant tube (RFIT), a set-valued function enclosing all possible state trajectories under a given feedback control law, regardless of the uncertainty realization. In this project, the main idea is to characterize RFITs with polytopic cross-sections via a min-max differential inequality for their support functions. This result leads to a conservative but tractable polytopic tube MPC formulation, which can be solved using existing optimal control solvers.  [Paper]  [Slides]  [Video]

---

FaSTrack: a Modular Framework for Real-Time Motion Planning and Guaranteed Safe Tracking

IEEE Transactions on Automatic Control (in revision), 2018

Real-time and guaranteed safe trajectory planning is vital to many applications of autonomous systems. However, algorithms for real-time trajectory planning typically sacrifice robustness to achieve computation speed. Alternatively, provably safe trajectory planning tends to be computationally intensive. We propose FaSTrack, Fast and Safe Tracking, to allow for real-time robust planning of dynamical systems. Within this framework, nonlinear model predictive control (NMPC) is used to achieve real-time trajectory optimization using a simplified and efficient planning model of the autonomous system. The plan is tracked using the autonomous system, represented by a more realistic and higher-dimensional tracking model. Leveraging Hamilton-Jacobi reachability, we precompute the worst-case tracking error due to model-mismatch between the tracking and planning models, as well as due to external disturbances (e.g. wind). This framework allows for fast online planning using the planning model, with guaranteed safe real-time tracking of the plan using the TEB and optimal tracking controller.  [Paper]  [Slides]

---

Plug and Play Distributed MPC for Heavy Duty Vehicle Platooning

57th IEEE Conference on Decision and Control, 2018, Miami Beach

Heavy duty vehicle (HDV) platooning has been widely accepted as a solution to reduce fuel consumption and traffic congestion. However, control strategy for HDV platoons interacting with other vehicles is not yet well established. This work presents a new framework for handling passenger vehicles (PV) plugging in or out of an HDV platoon. First, basic cruising control of the platoon is achieved by a distributed MPC scheme. Second, redesign of controllers ensures the stability of closed loop Plug and Play (P&P) operation. Finally, a transition phase to steady-states guarantees the feasibility of the newly synthesized controllers. For the plug-in case, we propose a novel Formation Coordinator that determines the optimal location at which the redesigned controller has the best initial feasibility.  [Paper]  [Slides]

---

Real-Time Tube MPC Applied to a 10-State Quadrotor Model

2018 American Control Conference, Milwaukee

This project is on a real-time implementation of tube MPC for nonlinear input-affine systems. This is achieved by combining recent theoretical and practical advances on the construction of forward invariant tubes with state-of-the-art algorithms for nonlinear MPC, such as the real-time iteration scheme. The focus is on presenting these ideas in a tutorial style, using a 10-state quadrotor model as an example. The controller is implemented using the automatic code generation capabilities of ACADO Toolkit. Numerical experiments show that the tube MPC scheme can achieve run-times in the lower millisecond range.  [Paper]  [Slides]  [Errata]

Notes on Control Theory

• Notes on Interpretable Robust Control and System Level Synthesis for ESE 680-004: Learning and Control, a course taught by Prof. Nikolai Matni at UPenn in Fall 2019.
• Notes on Jordan Chain written by me (as part of course materials for EE221A Linear System Theory taught by Prof. Claire Tomlin at UC Berkeley) [alternative link]