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 Experiment Videos

From IMU Streams to Real-Time Decisions: Past-Only Next-Window Badminton Action Prediction.

Qinglin Zhu1, Jiao Wang2, Bin Guo1

  • 1College of Electrical Engineering, Sichuan University, Chengdu 610065, China.

Sensors (Basel, Switzerland)
|June 26, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

End Point Prediction: Gran Plot01:07

End Point Prediction: Gran Plot

A Gran plot is used to predict the equivalence volume or endpoint of a potentiometric or acid-base titration without reaching the endpoint. Typically, titration data is collected as a function of the titrant's volume up to a point less than the equivalence volume and then transformed into a linear format. The straight line is extended to the x-axis, indicating the necessary titrant volume to achieve the equivalence point.
For potentiometric titration, the Gran plot is created by plotting the...

You might also read

Related Articles

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

Sort by
Same author

CogMamba: Multi-Task Driver Cognitive Load and Physiological Non-Contact Estimation with Multimodal Facial Features.

Sensors (Basel, Switzerland)·2025
Same author

Antiarrhythmic effects of ginsenoside Rg2 on calcium chloride-induced arrhythmias without oral toxicity.

Journal of ginseng research·2020
Same author

Association between acute severe mercury poisoning and multiple organ failure.

American journal of translational research·2020
Same author

Effects of co-loading of polyethylene microplastics and ciprofloxacin on the antibiotic degradation efficiency and microbial community structure in soil.

The Science of the total environment·2020
Same author

Analysis of regional river blocking by debris flows in response to climate change.

The Science of the total environment·2020
Same author

Lipid Dynamics, Identification, and Expression Patterns of Fatty Acid Synthase Genes in an Endoparasitoid, <i>Meteorus pulchricornis</i> (Hymenoptera: Braconidae).

International journal of molecular sciences·2020

This study presents a real-time badminton action prediction system using wearable IMU data. The BiLSTM + MHSA model achieves high accuracy and speed, enabling live coaching and tactical analysis.

Area of Science:

  • Sports Science
  • Wearable Technology
  • Machine Learning

Background:

  • Real-time action prediction is crucial for sports analytics.
  • Wearable Inertial Measurement Units (IMUs) offer rich data for movement analysis.
  • Class imbalance in continuous data streams poses a significant challenge.

Purpose of the Study:

  • To develop a real-time badminton action prediction system.
  • To predict upcoming actions using only past sensor data (causal inference).
  • To address class imbalance in continuous IMU data streams.

Main Methods:

  • Utilized window-level downsampling for dominant background class.
  • Applied Principal Component Analysis (PCA) for feature compression.
  • Employed a BiLSTM + MHSA model for temporal action prediction.
Keywords:
BiLSTMaction predictionbadmintoninertial measurement unit (IMU)principal component analysis (PCA)real-time inferenceself-attentiontime-series classificationwearable sensing

Related Experiment Videos

  • Evaluated the pipeline using a hop-based streaming protocol.

    • Main Results:

    • Achieved high recognition performance with 96.36% test accuracy and 95.82% Macro-F1 score.
    • Demonstrated real-time deployability, reaching 58.20 windows/s.
    • Exceeded the real-time requirement of 10 windows/s by 5.82 times.
    • System performance was validated on a Windows PC with an NVIDIA RTX 3080 GPU.

    Conclusions:

    • The developed system enables accurate and low-latency badminton action prediction.
    • The approach effectively handles class imbalance and real-time processing constraints.
    • Results support applications in live coaching feedback and tactical analytics.