You are here

Joshua Vaughan

Associate Professor

Ph.D. Mechanical Engineering, Georgia Institute of Technology, 2008
M.S. Mechanical Engineering, Georgia Institute of Technology, 2004
B.S. Physics with Honors Hampden-Sydney College, Hampden-Sydney, Virginia, 2002
B.S. Applied Mathematics Hampden-Sydney College, Hampden-Sydney, Virginia, 2002

Contact Information:

P.O. Box 43678
Rougeou Hall, Room 225
Phone: (337) 482-1207


Courses Taught:

MCHE 201 - Introduction to Engineering Design
MCHE 475 - Control Systems
MCHE 485 - Mechanical Vibrations
MCHE 513 - Intermediate Dynamics

Research Interests:

Improving Control of Human-Operated Systems
The efficiency and safety of many systems is still limited by our ability to control unwanted system dynamics, such as vibration. Furthermore, many, if not most, of these systems are controlled by human operators. However, the interaction between the operators and the control system is not well understood. New designs resulting from improved understanding of these interactions could have an enormous positive economic impact. Toward this end, we pursue not just better control algorithms but also investigate their interaction with human operators. The goal of this work is to establish known principles to be used during a concurrent controller and user-interface design process.

Integration of traditional control algorithms and machine-learning
Another area of interest and current research work is on improving our understanding of how recent advances in machine learning, and specifically reinforcement learning, enhance control of human-operated systems. The overwhelming majority of the work in reinforcement learning and other machine learning methods toward control of physical systems focuses on complete automation of the task; little-to-no consideration is given to integrating these methods with more-traditional control algorithms or to the inclusion of a human operator. This leaves a significant knowledge gap in the understanding of these algorithms, which we aim to fill.

Autonomous Maritime Systems
We are also working to improve the modeling and control of maritime systems, with a current emphasis on improved modeling, trajectory tracking, and obstacle avoidance for Autonomous Surface Vehicles (ASVs). In the initial work in this area, the majority of the object identification and segmentation was conducted through traditional machine-vision techniques. However, recently, these methods have been supplemented with a combination of machine learning algorithms. Even the relatively simple, initial implementations of these methods led to dramatic improvements in recognition of buoys, docks, and other vessels. Integration of these methods of obstacle identification into the vessel control system have led to similar performance gains.

Related to this work is the Maritime RobotX Challenge, for which teams of students develop the propulsion and control system for an autonomous boat. We first attended the contest in December 2016 and plan to enter the contest in future iterations of it.

Honors and Awards:

  • 2016 – 2017 University of Louisiana at Lafayette Innovator Award Winner
  • 2014 – 2016 University of Louisiana at Lafayette Rising Star Award Winner
  • 2014 – 2016 University of Louisiana at Lafayette Innovator Award Winner
  • 2015 Young Researcher of the Year – UL Lafayette College of Engineering
  • 2013 – 2014 University of Louisiana at Lafayette Rising Star Award Winner

Selected Publications:

D. Newman, S.-W. Hong, and J. Vaughan, The design of input shapers which eliminate nonzero initial conditions, Journal of Dynamic Systems, Measurement, and Control, 2018. and Control, vol. 140, no. 10, p. 101005, 2018.

A. Dhanda, J. Vaughan, and W. Singhose, Time-optimal and near time-optimal vibration reduction control for non-zero initial conditions, Journal of Dynamic Systems, Measurement, and Control, vol. 138, pp. 041006–041006, 02 2016.

J. Vaughan, A. Yano, and W. Singhose, Robust negative input shapers for vibration suppression, Journal of Dynamic Systems, Measurement, and Control, vol. 131, no. 3, p. 031014, 2009.

J. Vaughan and M. S. Rahman, Method for near-realtime workspace mapping, November 14, 2017. US Patent 9,818,198.