Gregory C. Walsh

Stability analysis of networked control systems

We introduce a control network protocol, try-once-discard (TOD), for networked control systems (NCS), and provide, for the first time, an analytic proof of global exponential stability for both the new protocol and the commonly used statically scheduled access methods. Controllers are designed without regarding the presence of the network in the feedback loop, so consequently many controller design techniques may be employed. The performance of the new network protocol and the statically scheduled protocols are compared in simulations.

Asymptotic behavior of nonlinear networked control systems

The defining characteristic of a networked control system (NCS) is having a feedback loop that passes through a local area computer network. Our two-step design approach includes using standard control methodologies and choosing the network protocol and bandwidth in order to ensure important closed-loop properties are preserved when a computer network is inserted into the feedback loop. For sufficiently high data rates, global exponential stability is preserved. Simulations are included to demonstrate the theoretical result.

Male displays adjusted to female's response.

Models of sexual selection generally assume that behavioural courtship displays reflect intrinsic male qualities such as condition, and that males display with maximum intensity to attract females to mate. Here we use robotic females in a field experiment to demonstrate that male satin bowerbirds (Ptilonorhynchus violaceus) do not always display at maximum intensity — rather, successful males modulate their displays in response to signals from females. Our results indicate that sexual selection may favour those males that can produce intense displays but which know how to modify these according to the female response.

Kinematics of a novel three DOF translational platform

A novel 3-DOF parallel manipulator is presented that employs only revolute joints and constrains the manipulator output to translational motion. Closed-form solutions are developed for both the inverse and forward kinematics. The inverse kinematics produces four solutions for each leg of the manipulator. In general, the four solutions are realized in only two unique leg configurations. The forward kinematic solution is reduced to a quadratic equation. So that in general, there are two poses the manipulator can assume for a given set of input joint angles.

Optimization of a three DOF translational platform for well-conditioned workspace

Two optimization studies on the design of a three degree of freedom translational parallel platform are conducted and the results are compared. The objective function of the first study maximizes total volume of the manipulator workspace without regard to the quality of the workspace. The second study optimizes the total volume of well conditioned workspace by maximizing a global condition index. The global condition index is a function of the condition number of the Jacobian matrix, providing a means of measuring the amplification error between the actuators and the end effector. Both objective functions involve an integration over the workspace of the manipulator. This integral is approximated using the Monte Carlo method.

Asymptotic behavior of networked control systems

The defining characteristic of a networked control system (NCS) is having a feedback loop that passes through a local area computer network. This paper considers nonlinear systems controlled in this manner, and demonstrates that for sufficiently high transmission rates, the network may be considered transparent. Three methods of scheduling data packets are compared: a static scheduler (token ring), the try-once-discard (maximum-error-first) scheduler with continuous priority levels, and the try-once-discard scheduler with discrete priority levels. The third method is of particular interest when only a small number of bits are available for collision resolution. Asymptotic stability is guaranteed in the first two cases, and ultimate uniform boundedness in the third. In the final section, simulations demonstrate the theoretical results. The contributions of this paper are two-fold: first, it extends the earlier results on NCS to nonlinear systems, and second, it allows for finite word-length message identifiers.

Brief Error encoding algorithms for networked control systems

A networked control system is characterized by having a feedback loop closed through a local area network. This paper considers methods for scheduling the use of the network to guarantee both stability and controller performance. We propose and validate algorithms for choosing message identifiers for dynamically scheduled networked control systems. Two schemes for selecting priority levels are proposed: a fixed arbitrary grid and an auto-scaling grid. We prove that the system is uniformly ultimately bounded in the case of the fixed encoding scheme, and asymptotically stable with auto-scaling. An inverted pendulum is used to illustrate the encoding methods.

On the application of multi-parameter extremum seeking control

In many control applications the best set point for operation is not known a priori, and yet, controller design tools tend to focus on regulation to known set points and reference trajectories. This paper presents and analytically verifies techniques for online optimization of set points. The main contributions of this paper include a new control law for multiple parameter set-points and a proof of exponential stability for the averaged system. Methodologies for implementing this control law in a discrete time setting are discussed. The application of this new algorithm to a magnetically suspended flywheel is detailed. The extremum seeking controller is used to minimize vibrations and the results of twenty runs compared with the expected physical model of the process. The fit is poor enough to justify the use of the online optimizer proposed.

Error encoding algorithms for networked control systems

We propose and validate algorithms for choosing finite word length priorities for dynamically scheduled networked control systems. Two schemes for selecting priority levels are studied, the first with a fixed arbitrary grid and the second with an auto-scaling grid. We prove that the system is uniformly and ultimately bounded in the case of the static encoding, and asymptotically stable with the auto-scaling methodology. Simulations of an inverted pendulum controller using these schemes are compared against a controller with point-to-point wiring.

Clear Panels: a technique to design mobile application interactivity

We introduce a design technique, Clear Panels, to design interactive mobile device applications. Using mixed-fidelity prototyping, a combination of low- and high-tech materials, participants refine multiple aspects of a mobile applications design. Clear Panels supports writing and sketching via a transparent overlay affixed atop a mobile device screen. It enables design partners to refine their gesture-based interactions on actual devices. The technique has been successfully implemented in the design of childrens mobile applications. The technique leverages and extends longstanding interaction design methods to include mobile and hand-held technologies. Importantly, we show it is effective in raising participants awareness of key mobile application design issues without constraining their creativity.