TLIO: Tight Learned Inertial Odometry

Liu, Wenxin; Caruso, David R.; Ilg, Eddy; Dong, Jing; Mourikis, Anastasios I.; Daniilidis, Kostas; Engel, Jakob

doi:10.1109/lra.2020.3007421

Cited by 151 publications

(102 citation statements)

References 15 publications

Supporting

Mentioning

101

Contrasting

Unclassified

Order By: Relevance

“…For metrics denoted with a dagger ( ) we excluded samples where the gold-standard motion speed was below 0.1 m/s. Performance metric Unit Description Formula TLE % Relative difference between trajectory length estimated by T265 and gold standard length ATE m RMS of distances between position estimated by T265 and OTS across the full trajectory 22 RTE m RMS of relative distance between position estimated by T265 and OTS over a window of k samples 22 TDr % Distance between final position estimates of a trajectory relative to trajectory length 22 GDE ° Mean angle between representation of gravity vectors in the respective calibrated frames , ° Roll and pitch angle difference , AYE ° RMS of yaw angle difference across full trajectory 22 RYE ° RMS of yaw angle difference over a window of k samples 22 YDr °/h Final yaw angle difference relative to trajectory duration T 22 …”

Section: Methodsmentioning

confidence: 99%

Evaluation of the Intel RealSense T265 for tracking natural human head motion

Hausamann

Sinnott

Däumer

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Accurate and robust tracking of natural human head motion in natural environments is important for a number of applications including virtual and augmented reality, clinical diagnostics, as well as basic scientific research. IMU provide a versatile solution for recording inertial data including linear acceleration and angular velocity, but reconstructing head position is difficult or impossible. This problem can be solved by incorporating visual data using a technique known as visual-inertial simultaneous localization and mapping (VI-SLAM). A recently released commercial solution, the Intel RealSense T265, uses a proprietary VI-SLAM algorithm to estimate linear and angular position and velocity, but the performance of this device for tracking of natural human head motion in natural environments has not yet been comprehensively evaluated against gold-standard methods. In this study, we used a wide range of metrics to evaluate the performance of the T265 with different walking speeds in different environments, both indoor and outdoor, against two gold-standard methods, an optical tracking system and a so-called perambulator. Overall, we find that performance of the T265 relative to these gold-standard methods is most accurate for slow to normal walking speeds in small- to medium-sized environments. The suitability of this device for future scientific studies depends on the application; data presented here can be useful in making that determination.

show abstract

Section: Methodsmentioning

confidence: 99%

Evaluation of the Intel RealSense T265 for tracking natural human head motion

Hausamann

Sinnott

Däumer

et al. 2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Such tracking usually yields more noisy positional estimates. Ongoing research is aimed at improving inertial tracking for the specific application of tracking human head position 20 .…”

Section: Discussionmentioning

confidence: 99%

“…Roll and pitch angle differenceα − α,β − β Absolute yaw error (AYE)°RMS of yaw angle difference across full trajectory 20 1…”

Section: Performance Metricsmentioning

confidence: 99%

Validation of the Intel RealSense T265 for Tracking Natural Human Head Motion

Hausamann

Sinnott

Däumer

et al. 2021

Preprint

View full text Add to dashboard Cite

Accurate and robust tracking of natural human head motion in natural environments is important for a number of applications including virtual and augmented reality, clinical diagnostics, as well as basic scientific research. Inertial measurement units (IMUs) provide a versatile solution for recording inertial data including linear acceleration and angular velocity, but reconstructing head position is difficult or impossible. This problem can be solved by incorporating visual data using a technique known as visual-inertial simultaneous localization and mapping (VI-SLAM). A recently released commercial solution, the Intel RealSense T265, uses VI-SLAM to estimate linear and angular position and velocity, but the performance of this device for tracking of natural human head motion in natural environments has not yet been comprehensively evaluated against gold-standard methods. In this study, we used a wide range of metrics to evaluate the performance of the T265 with different walking speeds in different environments, both indoor and outdoor, against two gold-standard methods, an optical tracking system and a so-called perambulator. Overall, we find that performance of the T265 relative to these gold-standard methods is most accurate for slow to normal walking speeds in small- to medium-sized environments. The suitability of this device for future scientific studies depends on the application; data presented here will be useful in making that determination.

show abstract

“…In another work, CNN was used to constrain the velocity in the Kalman filtering for PDR [50]. TLIO used residual network (ResNet) to compute displacements and their uncertainty and then integrated them into EKF [51]. The second approach relies on DNN to compute the odometry in an end-to-end manner without filtering-based approaches.…”

Section: Related Workmentioning

confidence: 99%

Understanding the Behavior of Data-Driven Inertial Odometry With Kinematics-Mimicking Deep Neural Network

2021

View full text Add to dashboard Cite

In navigation, deep learning for inertial odometry (IO) has recently been investigated using data from a low-cost IMU only. The measurement of noise, bias, and some errors from which IO suffers is estimated with a deep neural network (DNN) to achieve more accurate pose estimation. While numerous studies on the subject highlighted the performances of their approach, the behavior of data-driven IO with DNN has not been clarified. Therefore, this paper presents a quantitative analysis of kinematicsmimicking DNN-based IO from various aspects. First, the new network architecture is designed to mimic the kinematics and ensure comprehensive analyses. Next, the hyper-parameters of neural networks that are highly correlated to IO are identified. Besides, their role in the performances is investigated. In the evaluation, the analyses were conducted with publicly-available IO datasets for vehicles and drones. The results are introduced to highlight the remaining problems in IO and are considered a guideline to promote further research.

show abstract

TLIO: Tight Learned Inertial Odometry

Cited by 151 publications

References 15 publications

Evaluation of the Intel RealSense T265 for tracking natural human head motion

Evaluation of the Intel RealSense T265 for tracking natural human head motion

Validation of the Intel RealSense T265 for Tracking Natural Human Head Motion

Understanding the Behavior of Data-Driven Inertial Odometry With Kinematics-Mimicking Deep Neural Network

Contact Info

Product

Resources

About