Embedded Files
Research Team
Research Team
Faculty Members: Dr. Donna Scarborough & Dr. Michael Bailey-Van Kuren
Faculty Members: Dr. Donna Scarborough & Dr. Michael Bailey-Van Kuren
Student Researchers: Charlotte Cutright, Jillian Flinta, Ava Kahlig, Audrey Wendtland
Student Researchers: Charlotte Cutright, Jillian Flinta, Ava Kahlig, Audrey Wendtland
Introduction
Introduction
The Significance of the Transitional Cup (U.S patent no. 8432249, 2013)
The Significance of the Transitional Cup (U.S patent no. 8432249, 2013)
The goal of developing the transitional cup is to assist populations who experience dysphagia, or difficulty swallowing, while drinking liquids. The transitional cup regulates the rate at which fluid is released from the lid opening. The collected data contributes to future improvements in the cup design to improve patients’ quality of life by promoting independence, encouraging safe swallowing, and adjusting to patients’ varying swallow patterns.
Methodology & Lab Protocols
Methodology & Lab Protocols
Researcher 1 monitored the robotic arm’s movements during trials.
Researcher 2 initiated the arm’s movements and monitored data collection using the computer.
Protocols
Open the UFactory Studio Software that controls the xArm. The scale can be turned on which measures the fluid weight. Place the collection container on the scale and zero the scale to zero milliliters (mL). Measure the desired water volume (either 72 mL or 180 mL) and pour into the cup. Attach the lid with the sensor to the cup and attach the cup to the xArm. Set the UFactory software to the desired joint speed (either 25 deg/s, 20 deg/s, or 15 deg/s). Begin data collection on RealTerm software and ensure that there are no outside disturbances to the cup. Initiate pour mechanism of xArm using UFactory software. Monitor the trial and make note of any disruptions to the trial. Stop data collection on RealTerm software once xArm returns to resting position. Measure the volume of water dispelled from the cup as well as the residual volume of water remaining in the cup. Reset equipment and materials to prepare for the next trial.
Equipment
The design of the cup allowed the team to collect data through a sensor attached to the lid of the cup. The sensor collected data measuring acceleration and rotation on three axes. The data was saved on a desktop computer. A robotic xArm was utilized to standardize the pour mechanism. The cup attached to the xArm with a velcro strap. The team used a scale with a bowl to collect the dispelled water.
Figure 1a
Figure 1a
Displays the USB cord (x) connection to the microcontroller sensor which is secured to the lid.
Figure 1b
Figure 1b
Demonstrates the axes of the microcontroller, attached externally to the outside of the cup on the side of the lid (Fig 1a). The X-axis (yellow) runs from the attachment to the arm of the robot towards the researcher. The Y-axis (blue) is vertical from the bottom of the cup to the top. The Z-axis (orange) is perpendicular to the plane of the board.
Results
Results
During each robotic pour, the sensor provided 9,000-20,000 data points at a maximum rate of 100 hz. The sensor recorded data on 3 axis accelerations and 3 axis gyroscopes. Data was consistently pulled from roughly 700 data points from RealTerm (the software used to collect readings from the sensor). After analysis and graphical representation of all six sensor values, the research team focused on the acceleration data. The results table summarizes the sensor data related to the cup pouring during the flat segment of the robotic movement (image C above). The mean acceleration values were averaged over the steady state portion of the motion by identifying start and end data points within that range. The results collected in this study are pilot data due to the small sample size of experimental runs n=5-8 per case.
During each robotic pour, the sensor provided 9,000-20,000 data points at a maximum rate of 100 hz. The sensor recorded data on 3 axis accelerations and 3 axis gyroscopes. Data was consistently pulled from roughly 700 data points from RealTerm (the software used to collect readings from the sensor). After analysis and graphical representation of all six sensor values, the research team focused on the acceleration data. The results table summarizes the sensor data related to the cup pouring during the flat segment of the robotic movement (image C above). The mean acceleration values were averaged over the steady state portion of the motion by identifying start and end data points within that range. The results collected in this study are pilot data due to the small sample size of experimental runs n=5-8 per case.
Figure 2
Figure 2
Exhibits the data of the average acceleration of the cup along each axis during the pour cycle which is partitioned into five phases (A-E). These phases are depicted visually at corresponding points in the data on the y-axis (orange). At point A, the cup is resting on the table before the trial and the flat line represents no acceleration has occurring on the y-axis. Point B is the transition phase between the table and the complete pour position resulting in the steep slope of acceleration. Point C represents the fully tilted pour position (°) which lasted the duration of 30 seconds, hence no change in acceleration recorded. Point D represents the descent of the cup after pouring its contents and the steep slope of acceleration that happens. Point E is when the cup is returned to its resting place on the desk.
Figure 3 (a/b/c)
Figure 3 (a/b/c)
The range and distribution of the acceleration data for each case are depicted in Figure 3. The x-axis (3a) demonstrates little variation at low signal levels. The y-axis (3b) displays little variation near a value of -1 which corresponds to alignment with gravity (down). The z-axis (3c) shows consistency for low flow rates and shows an offset for higher flow and higher volume.
Average sensor acceleration (m/s^2) during the pour phase on the 3 axes of the cup mounted sensor. The x-axis (3a), y-axis (3b) and z-axis (3c) are shown for the four cases run in the experiment. Each box plot depicts the range with a line and the middle 50% with a box.
Discussion
Discussion
This study aimed to collect data on the effect that various volumes have on the data collected by the cup sensor. The data collected did not reflect this due to issues with ensuring a stable connection between the cup and the xArm at higher volumes. The x and y data obtained appears to be rotated 90 degrees from the manufacturer documentation. Given the orientation of the board on the cup, X values near -1 and y values near 0 would be expected during pouring. These axes would need to be calibrated for any eventual inclusion in the automated cup design. The z axis results identify a possible link between sensor data and the volume of liquid. This would allow the automated cup to incorporate this information into the cup controller’s estimate of the cup state as defined by the volume and angle of the cup. The data shows variability between trials due to the manual attachment of the cup to the robot. However, this has real world applications as the cup user will not lift and move the cup with a precise orientation. Whereas the study intended to identify an understanding of the cup contents using the on board sensor, the data analysis shows useful information related to the overall user movement of the cup in real time which can lead to a more accurate dynamic model that will refine the controlled flow of the transitional cup.
This study aimed to collect data on the effect that various volumes have on the data collected by the cup sensor. The data collected did not reflect this due to issues with ensuring a stable connection between the cup and the xArm at higher volumes. The x and y data obtained appears to be rotated 90 degrees from the manufacturer documentation. Given the orientation of the board on the cup, X values near -1 and y values near 0 would be expected during pouring. These axes would need to be calibrated for any eventual inclusion in the automated cup design. The z axis results identify a possible link between sensor data and the volume of liquid. This would allow the automated cup to incorporate this information into the cup controller’s estimate of the cup state as defined by the volume and angle of the cup. The data shows variability between trials due to the manual attachment of the cup to the robot. However, this has real world applications as the cup user will not lift and move the cup with a precise orientation. Whereas the study intended to identify an understanding of the cup contents using the on board sensor, the data analysis shows useful information related to the overall user movement of the cup in real time which can lead to a more accurate dynamic model that will refine the controlled flow of the transitional cup.
Future Directions
Future Directions
The data in this study was collected by a sensor taped to the cup lid. Eventually, this sensor will be embedded into the cup lid. The sensor readings will allow users to use the pinch valve mechanism to manipulate flow rate according to their personalized needs. In time, we would like to identify if varying volumes of water in the cup affect the prediction of movement and needed flow rate. Additionally, a future model will have reduced leakage and an air-tight lid, increasing the accuracy of the controlled flow rate.
The data in this study was collected by a sensor taped to the cup lid. Eventually, this sensor will be embedded into the cup lid. The sensor readings will allow users to use the pinch valve mechanism to manipulate flow rate according to their personalized needs. In time, we would like to identify if varying volumes of water in the cup affect the prediction of movement and needed flow rate. Additionally, a future model will have reduced leakage and an air-tight lid, increasing the accuracy of the controlled flow rate.
Page updated
Report abuse