CONIX Publication

Deep inertial sequence learning has shown promising odometric resolution over model-based approaches for trajectory estimation in GPS-denied environments. However, existing neural inertial dead-reckoning frameworks are not suitable for real-time deployment on ultra-resource-constrained devices due to substantial memory, power and compute bounds. Current deep inertial odometry techniques also suffer from gravity pollution, high-frequency inertial disturbances, varying sensor orientation, heading rate singularity and failure in altitude estimation. In this paper, we introduce TinyOdom, a framework for deploying neural inertial models on TinyML hardware. TinyOdom exploits hardware and quantization-aware Bayesian neural architecture search (NAS) and a temporal convolutional network (TCN) backbone to form the basis of scalable sub-meter inertial localization. In addition, we propose a magnetometer, physics and velocity-centric sequence learning formulation robust to preceding inertial perturbations. We also expand 2D sequence learning to 3D using a model-free barometric g-h filter robust to inertial and environmental variations. We evaluate TinyOdom for a wide spectrum of inertial odometry applications and target hardware against competing methods. Our framework yields accurate yet lightweight neural inertial localization models robust to sensory and ambient dynamics.