StepNet—Deep Learning Approaches for Step Length Estimation

Klein, Itzik; Asraf, Omri

doi:10.1109/access.2020.2993534

Cited by 59 publications

(32 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Specifically, HAR and SLR were shown to improve the accuracy of traditional pedestrian dead reckoning (PDR) by using it as a prior [21]- [25]. SLR was also shown to improve the performance of other navigation-related problems such as step length estimation [26]- [28] and adaptive attitude and heading reference system (AHRS) [29].…”

Section: Introductionmentioning

confidence: 99%

“…Recently, CNN and LSTM architectures were shown to improve SLR performance, compared to other learning-based approaches [30]. Methods coupling SLR with step length estimation proposed to use CNNs with or without LSTM [28], [30] or employed LSTM for SLR, similarly to previous works learning SLR for PDR [26]. Interestingly, CNN architectures with/without LSTMs yielded on-par performance, suggesting that LSTMs do not necessarily add an informative temporal aggregation, which is missing from CNNs, for this task [30].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Boosting Inertial-Based Human Activity Recognition With Transformers

Shavit

Klein

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

Activity recognition problems such as human activity recognition and smartphone location recognition can improve the accuracy of different navigation or healthcare tasks, which rely solely on inertial sensors. Current learning-based approaches for activity recognition from inertial data employ convolutional neural networks or long short term memory architectures. Recently, Transformers were shown to outperform these architectures for sequence analysis tasks. This work presents an activity recognition model based on Transformers which offers an improved and general framework for learning activity recognition tasks. For evaluation purposes, several datasets, with more than 27 hours of inertial data recordings collected by 91 users, are employed. Those datasets represent different user activity scenarios with varying difficulty. The proposed approach consistently achieves better accuracy and generalizes better across all examined datasets and scenarios.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Boosting Inertial-Based Human Activity Recognition With Transformers

Shavit

Klein

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…This process depends on a relationship between gait characteristic or sensor information and the step length as a training stage. Recently, machine learning and deep learning approaches are proposed to estimate the step length accurately [27]- [29]. The stacked autoencoders are proposed to estimate the step length based on the smartphone sensors, including accelerometers and gyroscopes [27].…”

Section: Introductionmentioning

confidence: 99%

“…In [28], the authors present a combined long short-term memory (LSTM) and denoising autoencoders to estimate pedestrian's stride length. Another approach for step length estimation is proposed by applying a one-dimensional convolutional neural network in [29].…”

Section: Introductionmentioning

confidence: 99%

Walking Step Length Estimation Using Waist-Mounted Inertial Sensors With Known Total Walking Distance

Pham

Suh

2021

IEEE Access

View full text Add to dashboard Cite

This paper presents a new constrained optimization-based smoothing algorithm for walking step length estimation using waist-mounted inertial sensors, where the total walking distance is known. The walking trajectory is estimated by double integrating acceleration. Due to sensor noises, the walking step length estimation accuracy degrades as the walking distance becomes longer. To tackle this problem, we introduce a known distance straight-line walking trajectory constraint and a constant speed constraint to the smoothing algorithm. These constraints reduce the walking step estimation accuracy degradation even for long walking distance. Two experiments are conducted to evaluate the pedestrian trajectory and walking step length estimation accuracy. The accuracy of a 20 m walking trajectory estimation has been investigated in the first experiment. This experiment compares the estimated position and velocity with Lidar-based references. The second experiment is to demonstrate the usefulness of the proposed walking step length estimation method. The result shows that the average of mean relative errors is 0.6801% for three different walking speed levels. The proposed method can be applied to generate training data for walking step length estimation without requiring spatial infrastructure.

show abstract

“…Step detection: the pedestrian steps are detected using accelerometer measurements [20] 2) Step length estimation: the pedestrian step-length can be determined using several approaches such as empirical formulas [21], [22] or deep learning [23] 3) Heading determination: the user walking direction is obtained from the gyroscope and/or magnetometer readings [24], [25] 4) Position update: the current user position is found giving initial conditions, heading estimation and step length estimation. Thus, in the PDR approach, only the accelerometers readings are used to determine the step length, i.e., the user change in distance.…”

Section: Introductionmentioning

confidence: 99%