We live in an aging society, an issue that will be exacerbated in the coming decades, due to low birth rates and increasing life expectancy. With the decline in physical and cognitive functions with age, it is of the utmost importance to maintain regular physical activity, in order to preserve an individual's mobility, motor capabilities and coordination. Within this context, this paper describes the development of a wireless sensor network and its application in a human motion capture system based on wearable inertial and magnetic sensors. The goal is to enable, through continuous real-time monitoring, the creation of a personalized home-based rehabilitation system for the elderly population and/or injured people. Within this system, the user can benefit from an assisted mode, in which their movements can be compared to a reference motion model of the same movements, resulting in visual feedback alerts given by the application. This motion model can be created previously, in a 'learning phase', under supervision of a caregiver.