|The Automatic Measurement of Behavior Using Sensor Arrays
|Tuesday, May 27, 2014
|10:00 AM–10:50 AM
|W184d (McCormick Place Convention Center)
|Area: PRA/EAB; Domain: Applied Research
|Chair: Nathan Blenkush (Judge Rotenberg Educational Center)
|Discussant: Ron Van Houten (Western Michigan University)
|CE Instructor: Nathan Blenkush, Ph.D.
We describe sensors and methods capable of detecting the force, duration, frequency, interresponse time, and topography of a wide range of behaviors over long periods of time. The sensors collect data 300 times per minute on 19 different variables useful in describing behavior. Each sensor is equipped with Bluetooth technology that allows it to connect with a computer. In addition, each sensor has an internal rechargeable battery. Taken together, the described sensors and methods have the potential to provide a valid, reliable, and simple method for measuring various dimensions of a response in real time. In research, such measurement will eliminate or reduce the need for interobserver agreement, subjective likert rating scales, and session based investigation. Such a measure will also facilitate direct comparisons between pharmaceutical, behavioral, and other therapeutic interventions. When combined with the standard celeration chart, the sensors will allow practitioners to better assess the efficacy of their interventions.
An Introduction to IMU Sensors
|NATHAN BLENKUSH (Judge Rotenberg Educational Center)
In science, it is often the case that the development of new measurement techniques allows for a greater understanding of functional relationships between variables. In behavior analysis, rate of response and the cumulative response recorder is an example of one innovation that led to many discoveries. However, few advances in measurement have arisen subsequent to Skinner's description of the cumulative response recorder. Here, the inertial measurement unit (IMU) is described. Each IMU is equipped with a 3-axis accelerometer, gyroscope, magnetometer, and other measurement devices. Taken together, this technology allows one to measure the force and direction of various movements. In addition, the orientation of the sensor during various movements can easily be determined. Internal clocks allow one to identify when a behavior occurred. When combined with databases, individual responses can be counted and automatically displayed on standard charts. These features allow behavior analysts to measure behavior over long periods of time without the need for a human witness.
Measuring Simple and Complex Behaviors
|JOSEPH TACOSIK (Judge Rotenberg Educational Center)
The inertial measurement unit (IMU) is ideally suited to measure simple human responses. The acceleration, angle, and absolute position of a single sensor often yield a unique signature that can be converted to a frequency count. Here, we describe methodologies developed in order to count simple behaviors using IMU data. IMUs produce extraordinary amounts of data. Therefore, databases to filter and analyze the data are necessary. However, many behaviors are topographically complex and require multiple sensor readings which further complicate measurement. Fortunately, advances in motion capture technology using IMUs allow for real time representations of the human form across an X, Y, Z grid relative to time. We summarize the state of the art in using motion capture technology to represent human responding. Implications for applied behavior analysis and the experimental analysis of behavior are discussed.