• The Avdanced Control Systems Lab
  • Research....
  • ...technology....
  • ...education

The Advanced Control Systems Lab


The primary goal of the Advanced Control Systems Lab (ACSL) is to perform research on advanced topics in modern control theory, such as nonlinear robust control, optimal control theory, differential game theory, and nonlinear estimation. Our theoretical results are tested by designing control systems for autonomous unmanned aerial vehicles and lightweight robots.

ACSL is a unique research facility comprised of a 6400 square foot hangar, where Dr. L'Afflitto and his collaborators perform indoor operations involving ground and aerial robots. This facility is equipped with a state-of-the-art Vicon motion capture system to measure position, attitude, and velocity of the vehicles being tested. Our research partners and sponsors include, but are not limited to, NSF, DARPA, ARL through a CRADA, NAVAIR.

Research Thrusts

Robust Adaptive Control in Presence of Constraints

Design of robust adaptive controls for trajectory following in the presence of constraints on the state and control space. Our control laws are effective also in case the reference signals violate these constraints.

Fast Model Predictive Control

We are designing a fast model predictive control (MPC) algorithm to generate reference trajectories for nonlinear dynamical systems subject to constraints on the stete, the outpur, and the control input and affected by external disturbances. This algorithms will be employed on UAS flying in unstructured environment.

Analytical Modeling of Mechanical Systems

Overly accurate models significantly impact the computational cost of the nonlinear robust control laws. We are devising techniques to produce dynamical model that are sufficiently accurate and yet simple to design nonlinear robust control laws under stringent computational constraints.

Analytical modeling of UAS' thrust in proximity of hard surfaces

In proximity of hard surfaces, the thrust produced by UAS' propellers cannot be modeled as proportional to the square of the angular velocity. In this research, performed in collaboration with Dr. D. K. Walters, we model the thrust force of multi-rotor UAS close to walls, floors, ceilings, and in the presence of cross-wind. These models are used to improve robustness marging in our UAS autopilots.

Statistical analysis of UAS trajectories

The outcome of flight tests depends on numerous stochastic factors, which are out of our control, such as wind and sensors' noise. In this research, we employ statistical analysis of our experiments and analyze the quality of our control algorithms.

Research Group

Graduate Students

  • Robert Anderson -- ISE -- Ph.D. Student
  • Julius Marshall -- ISE -- Ph.D. Student
  • Karen Martinez Soto -- ISE -- M.S. Student

Undergraduate Students

  • Lauren Ingmire -- AE at the University of Oklahoma

Selected Alumni

Selected Technological Products

We believe that sharing knowledge and technology is fundamental for the progress of science and technology. In the following, you find a selection of the computer codes produced at ACSL that we are pleased to share with everyone.

In this YouTube video, we show the results of a numerical simulation, wherein we control a tilt-rotor quadcopter equipped with an actuated inverted pendulum. This vehicle's autopilot implements a robust model reference adaptive control law that verifies user-defined constraints on the trajectory tracking error and the adaptive gains at all time.

In this YouTube video, we show step by step how to build one of the custom-made quadcopters currently in use at the Advanced Control Systems Lab. All materials used are listed and all steps are thoroughly commented. For its tutorial style, this video is recommended for amateurs and non-experts.

In this YouTube video, we show the final resul of a senior-year capstone project at the ACSL. A robotic arm was integrated with a monocular camera so that it could autonomosly look for a bottle, grab it, and poor its content in a bowl. The arm has no prior knowledge of the bottle's position, weight, or shape. In future, this arm will be installed on a quadcopter.

This Simulink toolbox is a CAD-based simulator for a quadtoror landing on a ship at sea; The aircraft can be controlled by a 4-axis joystick. This toolbox allows to compare and constrasts the ability of an adaptive controller and a PID-based controller to help pilots landing quadrotors despite the ship's movement, the wind, and the failure of one propeller.

This Simulink toolbox provides a working example of how multiple AR Drone Parrot 2.0 quadrotors can perform autonomous trajectory tracking by using Matlab 2015a and a Vicon system for motion capture. To see this toolbox in action, see this YouTube video.

This Simulink toolbox allows to compare any two control algorithms for quadrotors simultaneously. The vehicles are not modeled by ordinary differential equations, but by mean of a CAD model. PID- and MRAC-based autopilots are provided as working examples.

This Simulink toolbox provides a working example of how an AR Drone Parrot 2.0 can perform autonomous trajectory tracking by using Matlab 2015a and a Vicon system for motion capture.

Open Positions

Ph.D. Student

I would like to collaborate with a motivated student, who has good analytical skills and a strong background in mathematics and/or linear dynamical systems. Master of science in aerospace, mechanical, electrical engineering, mathematics, or closely related fields are preferable but not required. Applications from US citizens or permanent residents is strongly encouraged.

Undergraduate Students

I would like to collaborate with a motivated undergraduate student, who is interested in the design of advanced control algorithms using numerical methods. Proficiency in C or C++ and Matlab is preferable. This project is not limited to aerospace or mechanical engineering students only.

Undergraduate Students

I would like to collaborate with a motivated student, who is interested in the numerical of adaptive control techniques on quadcopters for automatic flight control problems. Proficiency in C or C++ and Matlab is preferable.

Interested candidates are invited to contact me via email.