The developer created a DIY ski coaching system, G-Force Skiing, as an alternative to expensive subscription-based solutions like Carv, which can cost $249 per season. The project aims to provide performance metrics using affordable IMU sensors and a custom mobile application, avoiding the high costs associated with proprietary hardware and cloud-based subscriptions.
The system tracks approximately ten metrics per turn, including edge angle, g-force, turn shape, and symmetry. These metrics are combined into a single score, GF Score, which is normalized based on the user's skiing style. The software uses a deterministic approach based on biomechanical physics rather than black-box machine learning models, ensuring transparency in how feedback is generated.
Hardware development involved testing different sensors, eventually settling on the WT9011DCL module due to its stability and cost-effectiveness. The total cost to assemble a pair of sensors is approximately 5000 ₽. The system uses Bluetooth Low Energy for data transmission, with all signal processing and calibration performed within the mobile application.
Initial testing and algorithm calibration were conducted on ice skates, which serve as a physical proxy for skiing movements like carving and edge transitions. The application includes features for real-time audio feedback during training, GPS tracking for free riding, and video integration to visualize performance metrics directly over recorded footage.
The development process utilized AI-assisted coding, focusing on a rigorous workflow where every constant and algorithm was verified against real data rather than relying on estimations. The project is currently in the pre-field-test stage, with future plans to implement automatic snow condition detection and further refine metrics for specific techniques like snowplowing.
