Student Research into Closed-loop System

Date Published: 
November 21, 2014

UNC Charlotte electrcial engineering student, Steve Banasik, collaborated with researchers from the Institute of Psychology of the Polish Academy of Sciences (IP PAN) to develop a closed-loop system capable of measuring latency in computer-based psychometric tools.

“Mechafinger,” the latency device, provides more accurate measurements of perceptual response rate by detecting delays that occur in computer hardware and software during human-machine interaction. It is able to detect the amount of lag time present in computer-based machines at a temporal resolution in milliseconds.

Numerous researchers fail to account for the delays that occur between a computer’s thought and action, or the time it takes a computer to respond to human action. The research team realized the need for a device that accounts for such delays since many cognitive systems operate at sub -- 100 millisecond rates, but the occurrence of certain cognitive phenomena is, at times, measured as deviations in the tens of milliseconds.

The validity of research data is increased when computer delay times are accounted for, allowing for more precise measurements and the reduction and control of error stemming from human-computer interfaces. The error reduction will ultimately help computer-based technologies improve simulation of how humans think and act.

Banasik mentioned: “[Researchers] want to know the actual machine’s delay and how late the actual person’s reaction is in relation to the machine’s delay, so that eventually when they’re simulating (for example) cognitive systems they’ll have the exact timings of the person’s and the computer’s response and a more valid representation of actual brain functioning.”

“Mechafinger” can improve computer-based research tools used to study and stimulate the brain’s functions. The lessened lag time in computer response rates will help researchers to better understand the brain as a dynamic system in terms of the onset and cessation of specific functions.

As described by Steve Banasik, the device was built to measure latencies from the equipment involved in computer-based psychometric tests. Data gathered from these tests are error-adjusted due to knowledge of latencies in the equipment.

According to Banasik, “this is extremely important in the fields of cognitive psychology and cognitive neuroscience since many phenomena are measured via reaction time.”  The difference between the presence of these phenomena and their absence is often measured in the range of 10-100 milliseconds (such as electroencephalography components), in which precise control of stimuli presentation and response measurement is crucial for the validity of results.

By measuring latencies, the procedures used in such experiments can be calibrated to compensate for inaccuracies in hardware and software. “Mechafinger” can be paired with any push-button machine. The team hopes to present the technology to universities focused on neuroscience, once the prototype has been finalized.

Banasik is currently a senior majoring in Electrical Engineering with a concentration in Power Electronics. His interest in Energy and Power Electronics developed after he became a research assistant for Dr. Babak Parkhideh in the “Inverter Molecule” project sponsored by EPIC and Sinewatt’s Inc. He contributed to the project as a way to put his education into practice and believes the project will be impactful in his future career.

“Thanks to EPIC and the ongoing research in the Power Electronics lab, I gained knowledge and skills that helped develop the “Mechafinger” system,” Banasik stated.

Project Abstract:  The specification provided by the Institute of Psychology, Polish Academy of Sciences as a basis for this project called for a closed-loop system for the precise measurement of latencies stemming from computer hardware and software in the context of computer-based psychometric tools.  Sources of latency in such tools includes the latency and refresh rate of computer displays, the polling loop time associated with most common input devices, and delays caused by processes of the software being employed.  In order to address these factors, I have designed and implemented a low-cost solution that makes use of a high-speed solenoid actuator to strike the input surface of an input device.  Once the input is provided, the physical time between input and appearance of the expected event on screen is measured by photodiode as a function of changes in luminance.  These two devices are coordinated by a microcontroller which provides a timeline of events associated with the procedure.  In terms of psychometric stimuli, this can be thought of as the timeline between the start of a test trial (charge sent to actuator) and the appearance of expected stimuli (photodiode detects change).

Project Highlights:

  • Creation of closed-loop system for measurement of latency (both hardware and software bound) in computer-based psychometric tools

  • Implementation of solenoid actuator and photodiode to automatically begin an experimental procedure and then measure the actual time of stimuli presentation and response

  • The use of a microcontroller programmed to coordinate the actions of the solenoid and photodiode in order to provide a timeline of events at a high temporal resolution