A good standing posture and two types of bad postures

More than 15 percent of people at age 65 and older report hip pain, and the common causes of hip pain include hip fracture, osteoarthritis and strain of the muscles in the hip joint [1]. Studies have found that abnormal loading of the hip is highly related to osteoarthritis, which consequently leads to hip pain [2]. Even though hip pain is a chronic condition and requires long term treatment, prevention is relatively easy. It was found that by decreasing the hip extension, patients would experience an immediate reduction in the pain [2]. In other words, people are often unaware that they are standing in a poor position (figure 1). By standing properly, the risk of getting hip pain would largely decrease. Nowadays, people realize the importance of employing wearable technology to keep a healthy posture. However, preventing hip pain is not addressed in available products. Most posture-correction commercial products only focus on back pain problems. Yet, as obesity has kept increasing in the past decade, the hip pain is going to be one of the most severe arthritis problems in the future. Thus, developing a convenient and reliable wearable device for the hip problem is inevitable.

Problem Statement

By simply standing in a good posture, people will face a lower risk of getting chronic hip pain. However, the market ignores the need as it is not as prevalent as the back-pain problem. Thus, we would like to tackle this problem by designing and developing an infrastructure for standing posture correction. To be specific, it is an infrastructure primary for sway-back posture, as it requires the most hip-extension. This hip-pain specific infrastructure will be able to collect posture data, perform accurate posture recognition and provide real-time feedback to remind users. Moreover, it will allow physical therapists to collect & label data, for the future use of more accurate machine learning algorithms. As a result, our project will offer an affordable solution to help potential users avoid hip pain; and it provides an open platform for posture-recognition research in lab settings.


For this senior design project, we aim to develop the infrastructure of a wearable device for sway-back posture correction. This device will be able to collect data from a wireless IMU attached to a belt worn by the user on the waist, apply efficient algorithms to correctly determine if the user is standing in sway-back posture for 90% of the time, and give off a vibration & LED feedback when incorrect posture is assumed. Thus, the device will be able to train the user to avoid a sway-back position when standing.  


Specifically, the development of the infrastructure consists of the following objectives: 

  1. Employ signal processing techniques to filter IMU data and extract posture information
  2. Develop algorithms to differentiate sway-back posture from different standing positions, such as standing straight and sway-front using IMU data
  3. Build infrastructure: a user-friendly wearable device and an interface in Python that allows the user to label data, for potential application of machine learning algorithm
  4. Add calibration functionality that allows the algorithms to adapt to different users 
  5. Test the prototype to evaluate its performance: find out probability of false alarm