Sheng Cheng

alt text 

Master Student
Research Assistant
Department of Electrical and Computer Engineering,
Institute for Systems Research
Phone: +1 301 335 299 five
Email: cheng [@] terpmail [DOT] umd [DOT] edu

About me

Hello! Thank you for visiting my website. I'm a research assistant at the University of Maryland, College Park. I'm currently studying the optimal control strategy for a mobile agent to reach a target enclosed within a denied area. I'm working with Dr. Nuno Martins. My Curriculum Vitae can be found here.


Research Interests

Robotics; control theory; optimization; perception; SLAM

Current Project

altd text 

Reaching a Target Inside a Denied Area Using the Optimal Control Method
Duration: July 2016–present
Project description: By now, various types of sensors have greatly broadened the scope of information that can be measured or detected. Yet there are situations where sensors can't work, i.e., the sensors can't provide reasonable readings. Such environment is called the denied environment or the denied area. For example, indoor environment is the denied environment for GPS. For a control system, the quality of the control relies closely on the quality of measurements. And when sensors can't work, the quality of control deteriorates significantly. In this project, we investigate a vehicle maneuvering problem. We want to control a mobile agent to reach a target enclosed within a denied area, where the relative position and velocity to the target are unavailable to the mobile agent. We will propose a dual mode controller using the optimal control method. The dual mode controller will use a closed-loop strategy when the mobile agent is outside the denied area and switch to an open-loop strategy when the mobile agent is inside the denied area. An example scenario is shown in the left. A terrorist is hiding in a known location inside a building and we want to send a drone to eliminate the terrorist. The environment inside the building is a denied area to the drone, because it can't receive GPS location or velocity information of itself. The proposed controller will enable the drone to reach the target without using indoor localization techniques which cause additional time and energy consumption to the drone.

The results on the inner part can be found in slides and poster. The results on the outer part is in progress.

This work is supported by the Air Force Center of Excellence: Nature Inspired Flight Technologies and Ideas (NIFTI).


  1. S. Cheng, “Research on ping-pong balls collecting robot based on embedded vision processing system,” in Applied Mechanics and Materials, vol. 511, pp. 838–841, Trans Tech Publ, 2014.

  2. S. Cheng, L. Huang, and Y. Li, “Model predictive control of frog-leg manipulator,” Journal of Shandong University of Science and Technology (Natural Science), vol. 33, no. 1, pp. 104–110, 2014.

  3. H. Wang, Y. Ma, S. Cheng, Y. Li, X. Lu, and X. Su, “Accurate robot calibration by SAI method through visual measurement,” in Robotics and Biomimetics (ROBIO), 2013 IEEE International Conference on, pp. 2679–2684, IEEE, 2013.

Honors and Awards

Courses (graduate-level)

Teaching and Mentoring



3181 A.V. Williams Bldg.
University of Maryland
College Park, MD 20742