Deep Learning Based Speed Estimation for Constraining Strapdown Inertial Navigation on Smartphones

Abstract

Strapdown inertial navigation systems are sensitive to the quality of the data provided by the accelerometer and gyroscope. Low-grade IMUs in handheld smart-devices pose a problem for inertial odometry on these devices. We propose a scheme for constraining the inertial odometry problem by complementing non-linear state estimation by a CNN-based deep-learning model for inferring the momentary speed based on a window of IMU samples. We show the feasibility of the model using a wide range of data from an iPhone, and present proof-of-concept results for how the model can be combined with an inertial navigation system for three-dimensional inertial navigation.

RINS-W: Robust Inertial Navigation System on Wheels

“…We also emphasize the importance of (11) and (13) in the procedure, i.e. applying (25)- (26). For illustration, we consider sequence urban07 of [11], where the vehicle moves during 7 minutes without stop so that ZUPT may not be used.…”

“…A vector y n+1 is computed by stacking the pseudomeasurements of the detected motion profiles. Note that, if z VEL n+1 = 1 we do not consider (25)-(26) since (23) implies (25)- (26). If no specific motion is detected, the update step is skipped, otherwise we follow the IEKF methodology [9] and compute…”

Deep-Learning-Based Pedestrian Inertial Navigation: Methods, Data Set, and On-Device Inference

“…Other data-driven methods [10], [11] learn to predict velocities in order to constrain system error drift, and achieve competitive performance. These learning based models have been shown to outperform previous model-based approaches in terms of accuracy and robustness [9], [10], [11]. There is a growing interest in applying deep neural networks to learn motion from timeseries data, due to its potential for model-free generalisation.…”

Robust Data-Driven Zero-Velocity Detection for Foot-Mounted Inertial Navigation

“…Our motivation for the use of learning stems in part from other work that has successfully applied learning-based methods to process inertial data. For example, Hannink et al [21] presented a method to train a deep convolutional neural network (CNN) to predict human stride length from inertial measurements, Chen et al [22] used an LSTM network to directly estimate a trajectory from raw inertial data, and Cortés et al [23] trained a CNN to regress velocities from IMU outputs to improve trajectory estimates.…”