Siddharth P. Nagarkatti

Robust visual-servo control of robot manipulators in the presence of uncertainty

This paper considers the problem of position control of planar robot manipulators via visual servoing in the presence of uncertainty associated with the robot mechanical dynamics and the camera system for both fixed-camera and camera-in-hand configurations. Specifically, we first design a robust controller that compensates for uncertainty throughout the whole robot-camera system and ensures global uniformly ultimately bounded position tracking for the fixed-camera configuration. Under the same class of uncertainty, we then develop a setpoint controller for the camera-in-hand configuration that achieves global uniformly ultimately bounded regulation.

Adaptive Boundary Control of Out-of-Plane Cable Vibration

This paper develops active controllers for the out-of-plane vibration of a flexible cable using boundary actuators and sensors, An exact model knowledge controller exponentially stabilizes the cable displacement assuming known system parameters. An adaptive controller asymptotically stabilizes the cable displacement while compensating for parametric uncertainty in the actuator mass and cable tension. The performance of the controllers is experimentally demonstrated.

Real-time Linux Target: a MATLAB-based graphical control environment

Real-Time Linux Target (RTLT) is a real-time Linux-based software environment for data acquisition and control that is integrated with MATLAB/SIMULINK. RTLT is a low-cost, user-friendly graphical tool for the design and implementation of digital computer control programs and is the first of its kind available under the Linux operating system. RTLT eliminates redundant hardware by using only one central processing unit (CPU) to perform deterministic real-time control computations as well as general-purpose non-deterministic processing of the Linux kernel and user programs (such as the user-interface), RTLT uses off-the-shelf consumer grade computer hardware, low-cost motion control boards, and a free operating system (OS). The non-proprietary open architecture makes this system more flexible than the current state-of-the-art as it is possible to combine various types of sensors from different vendors as inputs to one control program. The usual learning curve associated with new software tools is virtually eliminated as a result of the complete integration with the design and analysis tools provided by MATLAB/SIMULINK.

Tension and Speed Regulation for Axially Moving Materials

During continuous manufacture of axially moving materials such as fiber, paper, foil, and film, accurate speed and tension control are essential. In this paper, control torques applied to rollers at the boundaries of an axially moving system regulate the material speed and tension using speed and tension sensors for each roller. Given a distributed parameter model, Lyapunov techniques are used to develop a model-based boundary control system that exponentially stabilizes the material tension and speed at desired setpoints and stabilizes longitudinal vibration. Experimental results compare the tension and speed setpoint regulation provided by the proposed control strategy with proportional plus integral speed control and proportional tension feedback.

Autobalancing DCAL controller for a rotating unbalanced disk

We design a control strategy for a rotating unbalanced disk. The control strategy is composed of a control torque and two control forces. The control strategy regulates disk displacement while ensuring that the unbalanced disk tracks a desired angular velocity trajectory. Specifically the control uses a desired compensation adaptation law (DCAL) and a gain adjusted forgetting factor to achieve exponential stability despite the lack of knowledge of the imbalance-related parameters provided a mild persistency of excitation condition is satisfied.

Boundary control of a two-dimensional flexible rotor

In this paper, we present the design of boundary controllers for a two-dimensional, spinning flexible rotor system. Specifically, we develop a model-based boundary controller which exponentially regulates the rotors displacement and the angular velocity tracking error, and an adaptive boundary controller which asymptotically achieves the same control objective while compensating for parametric uncertainty. As opposed to previous boundary control work, which focused on the velocity setpoint problem and placed restrictions on the magnitude of the desired angular velocity setpoint, the proposed control architecture achieves angular velocity tracking with no restrictions on the magnitude of the desired velocity trajectory.

Boundary control for a general class of nonlinear actuator-string systems

In this paper, we design a controller for a boundary actuated-string model which avoids: 1) use of the standard small amplitude simplifying assumption, 2) use of the constant tension assumption, and 3) neglection of the actuator dynamics. Specifically, we design a nonlinear, exact model knowledge boundary controller which asymptotically stabilizes the total energy of the actuator-string system. We then redesign the exact model knowledge boundary controller as an adaptive boundary controller which compensates for parametric uncertainty and asymptotically stabilizes the total energy of the system.

Boundary control for a general class of string models

We study the control of an undamped, nonlinear string model with actuator dynamics at the boundary. Specifically, we develop a boundary controller which asymptotically stabilizes the out-of-plane displacement. The performance of the controller is illustrated via dynamic simulation.

Boundary control of a flexible cable with actuator dynamics

We develop boundary controllers for a flexible cable with actuator dynamics at the boundary. Specifically, we develop an exact model knowledge controller which exponentially stabilizes the position of the cable and an adaptive controller which asymptotically stabilizes the position of the cable while compensating for parametric uncertainty. The performance of the controllers is illustrated via experimental results.

Backstepping boundary control for noise reduction in finite ducts

In this paper, we develop a boundary controller to actively cancel acoustic disturbances in an enclosed duct of finite length using a loudspeaker. Given a distributed parameter acoustic model and the actuator electrical dynamics at the boundary, Lyapunov-type arguments are used to design a model-based backstepping boundary control law that asymptotically regulates the noise within the duct. Dynamic simulation results demonstrate the improved performance achieved by the proposed active noise cancellation algorithm.