This project was done as part of the thesis for the electronic engineering program of the Universidad del Norte. The technologies that were used were Django for the backend, react native for the frontend and python for data processing.
The recognition of human activities is a field whose applications range from telemedicine, in remote patient care and monitoring, to monitoring sports and / or military development. Activity recognition systems can be classified according to the way body movement information is processed. Online systems are those that classify the activity as the person carries it out. Offline systems are those in which data is collected while the activity is taking place and are only processed after the respective action has finished. In the case of the previously mentioned applications, online systems are the ones that manage to supply the need for on-site feedback of what the person does. In particular, online activity recognition systems using wearable sensors require constant monitoring of movement signals. Such monitoring results in constant energy consumption due to the need to transmit the data to a processing unit or to processing on the same device. In both cases, the energy availability of the power supply is a determining factor in the performance of the device. Therefore, architectures have emerged in the literature that propose the change of wearable motion sensors for elements that generate energy from the movement of the human body and that in turn provide information that allows classifying activities. However, the execution of these architectures has been limited to the recognition of activities offline and they do not present data about the extension of the useful life of wearable batteries. Consequently, the realization of this project seeks to be a contribution to the state of the art attacking these two shortcomings in the aforementioned architectures, providing information on the performance of an online activity classifier based on energy generation patterns together with the energy contributions towards the motion monitoring device battery.
- Build a system that recharges a battery and identifies a set of human activities using patterns of generating a kinetic energy harvesting scheme.
- Build a kinetic energy harvesting system that can be attached to a wearable device.
- Use a wearable device connected via Bluetooth to a cell phone to classify activities on site.
- Develop a mobile application that allows viewing historical data of activities carried out and of the energy transferred to the battery.