iot-based-atlete-performance-assessment-solution-back

IoT-based Athlete Performance Assessment Solution

Helps coaches evaluate athletes’ performance based on sensor data and personalize workout plans

Our client is a European consumer electronics brand that manufactures an array of wearable sensors and devices for the sports industry. At the moment the company is working on an athlete performance assessment platform with AI capabilities.

Problem

The company turned to Softeq to create an athlete performance assessment platform for sports coaches.

The solution is intended to:

Capture, process, and visualize the data acquired from custom sensing devices attached to an athlete’s body. The devices incorporate heart rate and IMU sensors
Enable coaches to monitor and evaluate athletes’ performance during a training session

Technology-wise, the platform would incorporate a cross-platform mobile app, cloud-based back end, and web dashboard.

Solution

Sensor data acquisition

We built a cross-platform mobile app that acts as a hub between the sensing devices and a back end, where sensor data is stored and processed. The app functions as follows:

Connects to the sensors and IMUs via BLE
Captures the data generated by the heart-rate sensors and IMU components (gyroscope, accelerometer, magnetometer)
Securely transfers the information to a cloud-based back end (Microsoft Azure)

Athlete performance assessment

To help coaches evaluate athletes’ performance based on sensor data, we built a web dashboard featuring:

.Net back end synced with the Azure data storage solution
Angular front end that visualizes sensor data in real time

The dashboard allows coaches to create athlete profiles, view workout data, and comment on it. By evaluating an athlete’s heart rate, body position, and acceleration, a coach can tell whether motion and force exercises are performed properly. This helps trainers personalize workout plans and improve athletes’ results.

Result

Our customer considers automating the evaluation process by feeding the annotated workout data to a neural network. Once we gather enough performance data, we’ll proceed to train a custom Machine Learning model and deploy it in the cloud.