Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

GCNTrack: A Pig-Tracking Method Based on Skeleton Feature Similarity.

Animals : an open access journal from MDPI·2025
Same author

An Automatic Movement Monitoring Method for Group-Housed Pigs.

Animals : an open access journal from MDPI·2024
Same author

Robust Flow Estimation Algorithm of Multichannel Ultrasonic Flowmeter Based on Random Sampling Least Squares.

Sensors (Basel, Switzerland)·2022
Same author

Reliable Indoor Pseudolite Positioning Based on a Robust Estimation and Partial Ambiguity Resolution Method.

Sensors (Basel, Switzerland)·2019
Same author

A New Method of High-Precision Positioning for an Indoor Pseudolite without Using the Known Point Initialization.

Sensors (Basel, Switzerland)·2018
Same author

[Effects of fluvastatin on expression of p38 mitogen-activated protein kinase in renal tissue of chronic graft versus host disease lupus nephritis in mice].

Zhonghua yi xue za zhi·2014

Related Experiment Video

Updated: Dec 22, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.1K

Deep Learning-Based Human Activity Real-Time Recognition for Pedestrian Navigation.

Junhua Ye1, Xin Li1, Xiangdong Zhang1

  • 1College of Geology Engineering and Geomantic, Chang'an University, 710054 Xi'an , Shanxi, China.

Sensors (Basel, Switzerland)
|May 6, 2020
PubMed
Summary
This summary is machine-generated.

Deep learning with Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN) significantly improves real-time recognition of pedestrian activities and smartphone posture for enhanced navigation accuracy.

Keywords:
CNNLSTMdeep learningpedestrian navigationtensorflow

More Related Videos

Trajectory Data Analyses for Pedestrian Space-time Activity Study
16:14

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Published on: February 25, 2013

14.0K
Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment
06:49

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

9.2K

Related Experiment Videos

Last Updated: Dec 22, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.1K
Trajectory Data Analyses for Pedestrian Space-time Activity Study
16:14

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Published on: February 25, 2013

14.0K
Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment
06:49

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

9.2K

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Sensor Technology

Background:

  • Pedestrian navigation systems often rely on smartphones but struggle with real-time recognition of user activities and device posture.
  • Traditional Machine Learning (ML) methods exhibit limitations in accuracy and timing for these recognition tasks.

Purpose of the Study:

  • To develop and evaluate a real-time recognition scheme for comprehensive human activities using deep learning and Micro-Electro-Mechanical System (MEMS) sensors.
  • To improve the accuracy and timing of pedestrian motion mode and smartphone posture recognition compared to traditional ML methods.

Main Methods:

  • Designed and trained deep learning models, including Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN), using the Tensorflow framework.
  • Utilized MEMS sensor measurements for real-time data acquisition.
  • Compared the performance of deep learning models against traditional ML classifiers like Support Vector Machine (SVM) and Neural Network (NN).
  • Transformed the trained CNN model into a .tflite format for deployment on Android devices.

Main Results:

  • CNN achieved 91.92% accuracy for motion mode recognition, surpassing SVM (89.9%).
  • CNN achieved 95.55% accuracy for smartphone posture recognition, surpassing NN (81.60%).
  • Real-time comprehensive pedestrian activity recognition reached an overall accuracy of 89.39% on a smartphone.
  • Navigation tests showed a reduction in mean bias by over 1.1m, indicating improved positioning accuracy.

Conclusions:

  • Deep learning algorithms, particularly CNN and LSTM, offer significant improvements in recognizing pedestrian activities and smartphone posture.
  • The developed scheme is effective for real-time, comprehensive activity recognition, aiding pedestrian navigation.
  • Applying deep learning (DL) to pedestrian navigation demonstrably enhances positioning accuracy and is a promising research direction.