Developers: Alexandra Kanaris, Katherine Lee, and Zoë Roecker
Advisors: Darcy Cooper, PT
Supervising Professor: Kevin Caves
Abstract
Parkinson’s Disease (PD) is a neurodegenerative disease that affects more than ten million people worldwide [1]. One prevalent symptom of PD is the loss of postural awareness and tendency to slouch forward. Our client was looking for a way to monitor and correct his posture. We created a system composed of a small wearable accelerometer that measures the angle of bend at the collar and an iOS app that connects via low energy Bluetooth. It then signals the iPhone to alert when the angle of tilt exceeds 10 degrees. This device allows the user to correct his posture independently and discreetly, which was verified with user testing and validation as well as clinician approval.
Introduction
Parkinson’s disease is a degenerative disorder of the nervous system that can affect motor abilities through the presence of tremor, rigidity, slowness, sensory disruption, and postural instability [2]. Our client, Gregg, is an upper middle-aged man with slow-onset Parkinson’s disease and only has difficulty maintaining an upright posture. He has a tendency to lose sense of what is an upright posture, but is physically capable of adjusting to this position if told to do so. He has full cognition, is able to operate small handheld devices, and is technologically savvy.
Gregg is a businessman that spends a lot of time interacting with clients and must give presentations, so he does not want others to notice he is slouching. This slouch is characterized by a forward-flexed neck and rolled shoulders. He has previously tried brace-type devices [3], but has not had success due to skin irritation and adjusting to its presence, known as habituation. There are options available on the market that monitor posture and alert the wearer to a slouch (such as LumoLift [4]), but all of the devices on the market only use one mode of notification, and none of them use the phone as a source of notification. Gregg desires to use the device throughout the day and during a wide range of activities including standing, sitting, walking, eating and driving. Since Gregg is able to hold upright posture when alerted, the alerts must be variant so he is able to respond to the device for the entire day without habituation.
Project Goals
The goal of the project is to develop a system that monitors posture and notifies the client that he is slouching. This device must be comfortable to wear throughout the day, and maintain functionality for at least 15 hours on a charge. It must be inconspicuous and work for both sitting and standing positions. Furthermore, the client should be able to use the device independently. Ideally, the device will not falsely alert the user during activities such as raising a hand or bending over to pick up an object. Since this will be a reminder system, it is imperative to use variant alerts such as vibrations and text alerts to prevent habituation and ensure efficacy.
Design and Development
The basis of this device is the use of an accelerometer to measure the angle of tilt. The device is placed on the client’s back. When the client stands up against a wall and is in perfect posture, he pushes a button in the One Thing Straight app to calibrate the device to an “upright” setting. As he declines into a slouched posture, his back curves and the device measures a different angle of tilt. If this angle difference exceeds 10 degrees, the client is alerted by a random vibration pattern through a vibration in his phone. The threshold of 10 degrees was chosen through analysis of the curvature of Gregg’s spine in images of good and bad posture (Figure 1.) This analysis was done in consultation with Darcy Cooper, PT, DPT, a physical therapist at Duke Hospital, to ensure the correction met current clinical requirements.
Figure 1: Images depicting slouched (left) and upright (right) posture
The device connects with the iPhone wirelessly using Low Energy Bluetooth, and signals the phone to vibrate if the threshold is exceeded. The phone will not stop buzzing until the angle of tilt is back under the threshold value, indicating that posture has been corrected. Overall, the device consists of 4 main components: the accelerometer detection system, the vibration notification system, the calibration button, and the hardware enclosure.
- Posture Detection System
The posture detection system consists of a 3-axis accelerometer BMA250 mounted onto the Bean, as pointed out in Figure 1. An accelerometer is a device that measures acceleration, or movement, by reading the forces pulling on a multiple axis scale. It is mounted parallel to the circuit board and thus aligns with and measures the tilt in the client’s upper back. Every 3 seconds, the accelerometer measures the acceleration in all axes as the cosine of the current angle. Using this data, the microcontroller computes the angle between the current vector of the accelerometer and the reference vector (xo, yo, zo), which is defaulted to the upward vector (0, -1, 0). A diagram showing the relevant vectors are shown in Figure 2. The value is then compared to a threshold value for bad posture measured on the client with medical guidance.
As seen in Figure 3, the user interface includes a button labeled “Sit/Stand.” The user can select the setting based on what they are doing at the time. The “standing” setting uses a delay of 2 seconds so that a slouched posture is quickly sensing can be easily corrected while walking. The “sitting” setting uses a delay of ten seconds so that small desk movements such as reaching for the phone or grabbing a folder do not trigger the notifications.
- Vibration Notification System
The vibration notification is embedded as an iOS application on our client’s iPhone. When the computed angle exceeds the threshold for bad posture, the vibration system is activated to notify the client to straighten his posture. To prevent habituation, or getting used to the cue and thus not responding, randomized sound and vibration patterns were implemented. If the client’s posture is out of compliance for an extended period of time, the sound pattern will continually change to alert the client. Since the posture measurement is triggered every 2 seconds, the client has 2 seconds to correct his or her posture prior to additional vibration notification. Sounds were selected such that notifications sound like a normal iPhone alert to others in the room so no attention is drawn. If the user is in a situation where he would not like sound notifications, the “Sound” button can be pressed and only vibration notifications will occur.
- Calibration Button
Because acceleration values measured from an accelerometer tend to drift over time, a calibration system must be implemented to reset the reference, or correct, vector (xo, yo, zo). The final component of the device is thus a button used for accelerometer calibration. The button is also embedded in the iOS application on our client’s iPhone. The user interface is shown in Figure 3. The client is advised to calibrate the device on the daily basis while aligning his back against a wall. When the calibration button is pressed, the reference vector (xo, yo, zo) is set to the acceleration values measured at that moment.
- Enclosure
The final component of our device is a 3D-printed encasing custom-made to house two magnets that are 0.5-inches in diameter along with the Bean microcontroller, as shown in Figure 4. The enclosure can be easily tucked under the back collar of a dress shirt, and secured with another two magnets placed on the other side of the collar fold. The thin enclosure allows for the device to align well with the client’s back and thus accurately detect posture.
This device detects when the client assumes a stooped posture, and notifies him via vibration to straighten his posture. This design is concealable and is inconspicuous under his professional attire. It is functional throughout a range of the client’s daily activities, including walking, sitting, standing, and driving. Additionally, it contains a calibration button to further ensure robust posture detection over time.
Evaluation
The device was assessed for discreetness, ease of independent use, durability, and effectiveness as a posture sensor for everyday use and activities. It was evaluated against the design specifications created for the device, and underwent verification testing in the laboratory and validation testing with the client. The client used the device over the course of three days for validation. He affirmed that the device functioned during walking, sitting, and standing activities, and that he did not habituate to the alerts. He said he is excited about using the device everyday and can definitely see himself using the device for a long time. All potential hazards were assessed and either redesigned, guarded, or warned against. Survey results and a complete hazard analysis is provided in Appendix C.
Discussion and Conclusion
This device is a posture monitoring and alert system that is specifically designed to address the needs of the client. Through the utilization of the iPhone as a notification system, only Gregg is privy that he needs to correct his posture and has not simply received a text message or email. The notifications blend in to the environment to maintain privacy for the user. Additionally, the random nature of the sound alerts ensure Gregg will not habituate to the notifications and forget to straighten his posture. These claims are supported by the fact that survey results were all “Satisfactory” or “Highly Satisfactory.” Additionally, Darcy Cooper, our clinical advisor, said the device would accomplish the design specifications and was very impressed with the small and discreet design. Overall, the system succeeds in providing the client a subtle way to monitor and correct his own posture during the course of the day.
References
[1] Statistics on Parkinson’s. (2017). Parkinson’s Disease Foundation. http://www.pdf.org/parkinson_statistics
[2] Understanding Parkinson’s. (2017). http://www.parkinson.org/understanding-parkinsons.
[3] Postural Extension Kyphosis and Osteoporosis Back Brace. (2017). http://www.braceability.com/cybertech-postural-extension-kyphosis-brace?utm_source=google_shopping&gclid=Cj0KEQiA_KvEBRCtzNil4-KR-LIBEiQAmgekF2QsDdqvFKNct4VTInKB3VJv7sgUGuQObOERWHQCVFQaAhZY8P8HAQ
[4] Lumo Lift Posture Coach. (2016). http://www.lumobodytech.com
Acknowledgements
We would like to thank all of the people who made the creation and implementation of this device possible: Kevin Caves, Richard Goldberg, Darcy Cooper, PT, DPT, and our client.
Contact Information
Team Members:
Zoe Roecker, zoe.roecker@duke.edu, 262-354-4915
Katherine Lee (Kathy), katherine.lee3@duke.edu, 540-425-4853
Alexandra Kanaris (Alex), alexandra.kanaris@duke.edu, 847-3021694