This Student-Built Smart Bracelet Offers Round-the-Clock Monitoring for ALS, Parkinson's Patients

Students at the Polytechnic University of Milan have developed a wearable smart bracelet that, they say, can deliver better in-home monitoring of patients with Amyotrophic Lateral Sclerosis (ALS) or Parkinson's disease — by tracking their movement with an on-board accelerometer.

"This device integrates an accelerometer to detect and record the vibrations and movements of the patient, enabling precise and continuous collection of clinically relevant data," the students explain. "The bracelet is designed to capture information in both the time and frequency domains, providing a detailed view of variations in the patient's movements. These data are then wirelessly transmitted to a dedicated server for storage and analysis, ensuring secure and centralized information management."

The wearable itself is based on an Arduino-compatible Wemos development board, powered by an Espressif ESP8266 microcontroller, connected to a TDK InvenSense MPU-6050 accelerometer sensor. The two devices are powered by a 250mAh lithium-polymer battery. Everything is kept together inside a 3D-printed fire-retardant ABS case, with a separator board to keep the battery safe from damage — and mounting holes for a strap, allowing the device to be worn on the wrist like a watch.

"The main concept behind this system is that the wristwear continuously reads acceleration values whenever it is connected to a Wi-Fi network and a listening server is available," the students explain. "The server can be any device, such as a PC or a smartphone/tablet, equipped with a dedicated software. This software is designed to receive and store data in real-time from the sensor. It then carries out data analysis to detect and report any possible tremors. To optimize performance, the data analysis is conducted offline on an acquisition file covering a specific time duration."

The analysis is carried out using a Python notebook running on the host system, which delivered an 82 percent accuracy rate in identifying tremors — replaced in the project's final tests with a lighter-weight Python script offering a claimed 94 percent detection rate of real-world tremors. The results from this script are then turned into a spreadsheet and automatically emailed to the patient's doctor for further investigation. "These data are also accessible to the patient through a mobile application," the students add, "current conceptualized but not yet implemented."

The team's full project, including source code and report, has been uploaded by co-author Matilde Grandi to Hackaday.io.