Inertia-1: An Open Exploration to a Unified Motion Foundation Model

Inertia-1 is a unified motion foundation model trained on wearable sensor data that generalizes across different body placements, sensor types, and sampling rates without retraining. The model demonstrates that a single backbone can perform multiple tasks—including activity recognition, gait analysis, and disease prediction—while maintaining performance even when deployed on devices and body locations it wasn't originally trained on. This approach addresses the fragmented state of motion modeling by establishing one general representation instead of requiring task-specific models for different configurations.