This repository implements the DynaFuse framework for stress monitoring by dynamically fusing multimodal sensor data to optimize prediction accuracy and energy efficiency. By adjusting modality weights in real time based on input quality and energy constraints, the system effectively balances the trade-off between accuracy and resource consumption, outperforming traditional static fusion methods. The implementation processes data from wearable sensors such as PPG, EDA, and ECG, enabling efficient and adaptive stress monitoring. Derived from the IEEE Embedded Systems Letters paper by Alikhani et al. (Dec 2023) citeturn0file0, this project offers a compact, configurable platform ideal for real-time, resource-constrained health applications.