Gregory W. Neat

An EEG-based brain-computer interface for cursor control ☆

This study began development of a new communication and control modality for individuals with severe motor deficits. We trained normal subjects to use the 8-12 Hz mu rhythm recorded from the scalp over the central sulcus of one hemisphere to move a cursor from the center of a video screen to a target located at the top or bottom edge. Mu rhythm amplitude was assessed by on-line frequency analysis and translated into cursor movement: larger amplitudes moved the cursor up and smaller amplitudes moved it down. Over several weeks, subjects learned to change mu rhythm amplitude quickly and accurately, so that the cursor typically reached the target in 3 sec. The parameters that translated mu rhythm amplitudes into cursor movements were derived from evaluation of the distributions of amplitudes in response to top and bottom targets. The use of these distributions was a distinctive feature of this study and the key factor in its success. Refinements in training procedures and in the distribution-based method used to translate mu rhythm amplitudes into cursor movements should further improve this 1-dimensional control. Achievement of 2-dimensional control is under study. The mu rhythm may provide a significant new communication and control option for disabled individuals.

An EEG-based method for graded cursor control

Individuals were trained to modulate their EEGs in order to move a cursor on a video screen to intercept a moving target. EEG activity was recorded from the scalp over the central sulcus of the left hemisphere, and mu-rhythm amplitude was assessed three times per second by a fast Fourier transform. The cursor began at the midpoint of the right edge of the screen and moved up or down depending on mu-rhythm amplitude. A target of selected vertical length began at a random height on the left edge of the screen and moved horizontally across the screen in 8 sec. The subjects’ task was to move the cursor along the right edge of the screen so as to intercept the moving target. After several weeks of training, 3 of the 4 subjects were able to perform this task with significant success. On average, these 3 subjects reduced the vertical target-cursor distance to 54% of its initial value. These results indicate that the mu rhythm can be used to control graded cursor movement and are additional evidence that with further development it might provide a new means of communication and control for individuals with severe motor disabilities.

Asymptotically stable multiple-input multiple-output direct model reference adaptive controller for processes not necessarily satisfying a positive real constraint

This paper presents an asymptotic output tracking stability proof for a new modification to a direct model reference adaptive control procedure. This modification, which alleviates a very restrictive positive real constraint, greatly expands the class of processes that can now be controlled with zero output error. This paper presents illustrative examples demonstrating the utility of the algorithm.

Expert adaptive control for drug delivery systems

The goal of this investigation is the development of an adaptive drug delivery system for use in regulation of critical care patients suffering from cardiac failure. It is assumed that adaptive control algorithms combined with expert system techniques are necessary to maintain stable patient statues within narrow physiological bounds in the presence of large plant uncertainty, and a hybrid controller is presented. Its structure is adjusted by an expert system that attempts to select the best control scheme in accordance with the dynamic structure of the plant. An illustrative example is presented, demonstrating the improved performance provided by the hybrid control scheme.<<ETX>>

A hybrid adaptive approach for drug delivery systems

The goal of this work is to develop an expert, hierarchical control scheme that adapts its structure in arcordance with a broad range of plant variations and knowledge about the plant. The coarsest control, provided by a fuzzy controller, moves the state of the plant towards the set point by modeling the actions of the operator. A multiple model adaptive control procedure provides the next level of control by forcing the plants response to exist within a predetermined allowable range of specifications. The finest control is determined by a model reference adaptive controller that causes the plant to follow a desired reference model. The heuristics used to determine switching between different controllers is orchestrated by an expert system that bases its decisions on plant responses and controller parameter values. An illustrative blood pressure controller example is presented demonstrating the improved performance provided by this hybrid control system.

A hybrid adaptive control approach for drug delivery systems

The goal of this investigation is the development of an adaptive drug-delivery system for use in regulation of critical-care patients suffering from cardiac failure. Adaptive control algorithms are combined with expert-system techniques to maintain stable patient status within narrow physiological bounds despite large plant uncertainty. To this end, a hybrid controller is presented whose structure is adjusted by an expert system that attempts to match the best control scheme in accordance with the dynamic structure of the plant. An illustrative example is presented, demonstrating the improved performance provided by this hybrid control scheme.<<ETX>>

EEG-based Brain-to-computer Communication: System Description

This paper describes an adaptive decoding device designed to develop and evaluate EEG-based communication from the device user to a computer. The decoder transforms EEG activity into cursor movement. The operator, who is considered part of the device, defines the parameters governing this transformation and makes adjustments periodically as the user’s control of the EEG changes. The user watches a video terminal that displays the cursor and the stationary target. The resulting device-user interaction forms a complex dual control structure in which both elements (the device and the user) adapt to achieve the ultimate mutual goal of fast, accurate control of cursor movement. The hardware components are a Grass model 8 EEG polygraph, a Texas Instruments TMS320C25 based signal processing board, and an Intel 80286 based IBM PC/AT computer. This combination of the high speed TI processor with the AT host processor supports the custom software that will ultimately allow the adaptive device to be operator free. The paper presents results illustrating successful EEG-based cursor control with this system.

Hierarchical Adaptive Control System Design and Application

Abstract This paper presents a new adaptive hierarchical control technique that adjusts its structure in order to maintain good performance over a wide variation of plant operating conditions. In contrast to the conventional approach of designing a compensator based upon a single control algorithm, this method combines a range of control algorithms in a complementary manner, each of which makes different assumptions about the regulated process. The hierarchical control structure includes a fuzzy controller which requires little process knowledge in order to provide an effective control, a multiple model adaptive controller which provides credible performance when the model bank adequately spans the expected plant variations and the plant behaves reasonably close to the model on which the compensator design was based, and a direct model reference adaptive controller which assumes that the plant transfer function satisfies specific sufficiency conditions. This paper presents an application of the hierarchical control structure as a regulator for a drug delivery system.