1. Intra-fractional Movement Detection in Immobilized
Radiology Patients
Erik Zaro
Advisor: Dr. Farid Farahmand
Collaborator: Dr. Robert Woodburn, Kankakee Radiation Oncology
Introduction
Current methods for detection of minute movements in patients are very accurate
and precise. However, the price of these systems can average around $80,000
with installation costs, and additional maintenance costs. The goal of this project is
to create a system that meets the minimum thresholds for movement detection in
radiology treatments with a unit cost of under $150 U.S. The device uses a 9-
degree sensor system containing an accelerometer, gyroscope, and magnetometer
for real-time, accurate data gathering. The system will be integrated with current
immobilization techniques in order to minimize intrusion on existing procedures.
Detector Software Implementation
The detector software is hosted on the end-user terminal. This program allows the
user to manually input an angular threshold. Once this threshold is set, the program
will continuously read in serial data; If any axis experiences angular movement
greater than the threshold; the program will trigger an alert message box detailing
which axis experienced movement. Once a patient has been repositioned, the
message box can be closed and the program will begin waiting for another
threshold exceeding movement. The equations below demonstrate how angular
movement can be mapped an approximate physical distance.
Bluetooth Communications
9-DOF Sensor with IMU
Bluetooth communication is handled by a Bluetooth Mate Gold circuit. This circuit
handles serial input from the 9-DOF IMU and outputs a wireless Bluetooth serial
stream that can be received by a compatible, paired terminal.
The 9-DOF sensor chip comes with 3 sensors: an accelerometer, a gyroscope, and a
magnetometer, as well as an integrated microprocessing unit (IMU). This allows for
some of the data processing to be done locally on the chip. The IMU takes the
information from each of the 3 sensors and calculates a 3-axis value
(Yaw,Pitch,Roll) which represents the current angular position of the sensor board.
This data is then sent to a serial terminal; this data can be picked up through an
end-user serial connection and applied in a variety of ways.
Device enclosure and Mounting
The device enclosure was designed to snap into mounts present on a Precise Bite
Patient Re-Positioner, an existing immobilization device used to deter patient
movement. By replacing the Precise Bite top plate with the detection device, new
technology is seamlessly integrated with current techniques for a minimally intrusive
operation. It also places the sensor at an optimal position on the patient’s body,
reducing noise, and maximizing sensor potential. Enclosures were 3D printed from
Solidworks models designed by Aram Yegiazaryan.
Testing setup and trials
Sensitivity and accuracy tests were conducted on a variety of different magnitudes
of movement. A rotating angular table was used to test the device. The table would
be rotated a set number of degrees, and sensor data would be used to compare the
programmed movement against the measured movement.
Results
Testing of the device showed that the sensor could reliably detect movements up
to 0.3 degrees, or approximately 1 millimeter. The device is very accurate and
sensitive for minute movement detection. However it does not perform as well
with accurately measuring larger magnitude movements.
9-DOF sensor with IMU Bluetooth Mate Gold Module
Rotating table testing setup
Sensor output with movement detection software