About 65 million global disable population require a wheelchair to keep mobility. However, long-term depending on wheelchairs has intensified the burden on the social economy and brought enormous challenges to wheelchair users health. Physical issues such as body pain, pressure sore, muscle injury and weakness, and cardiopulmonary weakness happen quite often among wheelchair users because of various poor long-term habits including, improper sitting postures, prolonged sitting, upper extremity overuse, and lack of exercise engagement. To solve these problems, wheelchair users have to change their habits.
Self-tracking based on emerging sensing technologies combined with the personal informatics might be a promising way to motivate behavior change. Such products have been applied in sports and monitoring chronic diseases such as migraines, diabetes, and allergies. The abundance of commercial smart trackers can also prove that self-tracking is competent to help users to self-understand, make plans, and finally, change habits. However, there are no available self-logging products specifically designed for wheelchair users. Furthermore, because of the limitation of their bodies and different needs from regular people, wheelchair users can hardly find suitable commercial products.
To solve these problems, a smart connected wheelchair system is designed which applies emerging self-tracking technologies and machine learning. The wheelchair can monitor postures, sedentary time, upper extremity muscle load, exercise level, and energy expenditure. The App associated can record and report this data for users' better self-understanding. Various strategies for motivating behavior changing, including goal-setting, reward, self-monitoring, and sharing, are also integrated into this system. In this way, wheelchair users can make new plans according to their data and keep moving toward a healthier lifestyle.
An agile design approach in combination with technology-driven rapid prototyping, which includes multiple iterations of technology experimentation, prototype development, and user test, was adopted throughout this project. In the prototyping phase, the concept was divided into three parts including sitting, muscle, and cardiopulmonary. Three generations of the prototype were then built, focusing on validating the sitting aspect of the concept from the technical perspective. By utilizing a simple machine learning algorithm, the function has been realized with a relatively satisfying accuracy. It turns out that the concept has potential to benefit wheelchair users in the future.