Machine learning technique-based diagnosis of wrist-radial pulse

A growing research on the nadi signal or the radial pulse has shown that physiological states and chronic illnesses are linked to each pressure change occurring in the arteries. The variation in the kapha, vata and pitta energies is associated with various reflexes that help our body to react to situations. The three energies together is known as Tridosha and each individual has a unique balance of tridosha. It requires great focus and accuracy to exactly determine the tridosha ratio by an Ayurvedic person. Owing to the advancement in technology in medical sciences, these signals can be collected, filtered, amplified and categorized with greater precision. The work proposed is used to implement the above-mentioned process in order to help patients self- diagnose the imbalance in the tridosha for diseases such as diabetes mellitus and blood pressure. The pulse signals are attained by pulse sensors positioned at the wrist. The signal passes through the Arduino where the analog counterpart are transformed to digital form. The converted signals are sent to the raspberry pi where the features are extracted from the signal and are compared with the model built based on a Machine Learning algorithm. The output is finally displayed indicating which doshas are not balanced and the illness which has occurred due to that imbalance.