Abstract:The International Journal of Robotics Research, Ahead of Print. In monocular visual-inertial navigation, it is desirable to initialize the system as quickly and robustly as possible. A state-of-the-art initialization method typically constructs a linear system to find a closed-form solution using the image features and inertial measurements and then refines the states with a nonlinear optimization. These methods generally require a few seconds of data, which however can be expedited (less than a second) by adding constraints from a robust but only up-to-scale monocular depth network in the nonlinear optimization. To further accelerate this process, in this work, we leverage the scale-less depth measurements instead in the linear initialization step that is performed prior to the nonlinear one, which only requires a single depth image for the first frame. Importantly, we show that the typical estimation of all feature states independently in the closed-form solution can be modeled as estimating only the scale and bias parameters of the learned depth map. As such, our formulation enables building a smaller minimal problem than the state of the art, which can be seamlessly integrated into RANSAC for robust estimation. Experiments show that our method has state-of-the-art initialization performance in simulation as well as on popular real-world datasets (TUM-VI, and EuRoC MAV). For the TUM-VI dataset in simulation as well as real-world, we demonstrate the superior initialization performance with only a 0.3 s window of data, which is the smallest ever reported, and validate that our method can initialize more often, robustly, and accurately in different challenging scenarios.

State initialization for recurrent neural network modeling of time-series data

Multistep Prediction of Dynamic Systems With Recurrent Neural Networks

Transient Modeling of High-Speed Links Using Transfer Learning-Based Neural Network Initialization

Initialization Approach for Nonlinear State-Space Identification via the Subspace Encoder Approach

How to Initialize your Network? Robust Initialization for WeightNorm & ResNets

Initialization of multilayer forecasting artifical neural networks

Dynamical Isometry and a Mean Field Theory of LSTMs and GRUs

Revising the Structure of Recurrent Neural Networks to Eliminate Numerical Derivatives in Forming Physics Informed Loss Terms with Respect to Time

Physics-Inspired Temporal Learning of Quadrotor Dynamics for Accurate Model Predictive Trajectory Tracking

Automatic Structure and Parameter Training Methods for Modeling of Mechanical System by Recurrent Neural Networks

Deep Cnn Identifier For Dynamic Modelling Of Unmanned Helicopter

Fast and robust learned single-view depth-aided monocular visual-inertial initialization

Deep State Space Recurrent Neural Networks for Time Series Forecasting

Stochastic Recurrent Neural Network for Multistep Time Series Forecasting

Improving performance of recurrent neural network with relu nonlinearity

Data-driven learning of non-autonomous systems

Application of Neural Network Based on Real-Time Recursive Learning and Kalman Filter in Flight Data Identification

Learning Interpretable Deep State Space Model for Probabilistic Time Series Forecasting

Flight Trajectory Prediction Based on Hybrid- Recurrent Networks

Deep Subspace Encoders for Nonlinear System Identification

Switching Recurrent Kalman Networks