Skip to content

Julian-BFacal/RacketSense

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Ideas comerciales

  • SONY: SSE-TN1W
  • Zepp Tennis 2

Github en que basearse

  1. Proxecto frances IMU
  2. Proxecto frances APP
  3. Dataset video
  4. Prueba datasets y modelos

Resumen Papers

Papers Tenis de Mesa

  1. Accurate Recognition of Player Identity and Stroke Performance in Table Tennis Using a Smart Wristband
  2. Action Recognition and Application of Table Tennis Training Based on IOT Perception

Papers Tenis

  1. An Approach to 3D Gyro Sensor Based Motion Analysis in Tennis Forehand Stroke
  2. An_embedded_6-axis_sensor_based_recognition_for_tennis_stroke
  3. Tennis Stroke Classification: ComparingWrist and Racket as IMU Sensor Position

7 Sensores de Gryo establecidos en diferentes puntos do corpo. A velocidade angular obtida polos sensores xiróscopos debe ser transformada en ángulos de roll-pitch-yaw.

Paper3modelo Microcontrolador de 6 eixes JY-61. Modulo de control STM32F405 por BLuettoth. Instalado en raqueta, realtime a mobil. (98% golpes, 96% efectos) 3 etapas:

  • Reconocemento de golpe (fluctuación de aceleración en ventanas).
  • Clasificación de golpe (forehand, backhand, saque) por aceleración e velocidad angular.
  • TopSpin o BackSpin rotación de pelota por velocidad angular.

Comparativa de aceleración maxima, desviación estandar e diferencia entre maximo e minimo. O mellor foi desviación estandar.

Smash e saque valores importantes en Y. Para diferenciar saque de smash, para o saque vai existir unha gran difrencia temporal entre o saque e o ultimo golpe ao contrario que no smash.

MIRAR REFERENCIAS INTERESANTES
Compara muñeca con raqueta. Para clasificar os tipos de golpes usa 4 modelos de machine learning. Métodos para detectar golpes:

  • Acelaración pico con threshold: 9g en x axis
  • Calculated accelaration de 3 o 8 gs
  • Falsos negativos fora con gaps de 1.25 segundos ou or alternatively gyroscope resultant values within 0.06 s on both sides of the acceleration maxima exceeding ±400 deg.

Machine Learning usados Naive Bayes , support vector machine (SVM), decision and classification tree(CT) random forest (RF) k-nearest neighbor (KNN), neural network (NN) and logistic regression(LG) Further models include AdaBoost and discriminant analysis.

Usan acceleration and angular velocity

Mellores resultados SVM linear con datos brutos de acelaración e giroscopio pasada por un PCA.

Outros

  1. Design and Performance Evaluation of An Arduino Based Activity Tracker

Ideas Traballo relacionado / Referencias

  1. L. Chen, J. Hoey, C. D. Nugent, D. J. Cook and Z. Yu, "Sensor-Based Activity Recognition," in IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 42, no. 6, pp. 790- 808, Nov. 2012.
  2. K. F. Li, A. M. Sevcenco and K. Takano, "Real-Time Classification of Sports Movement Using Adaptive Clustering," Complex, Intelligent and Software Intensive Systems (CISIS), 2012 Sixth International Conference on, Palermo, 2012, pp. 68-75.
  3. D. Conaghan, P. Kelly, and N. E. O’Connor. Game, shot and match: Event-based indexing of tennis. In 9th International Workshop on Content-Based Multimedia Indexing, pages 96–102, 2011.
  4. D. Connaghan, S. Hughes, G. May, K. O’Brien, P. Kelly, C. ´O Conaire, N. E. O’Connor, D. O’Gorman, G. Warrington, A. F. Smeaton, and N. Moyna. A sensing platform for physiological and contextual feedback to tennis athletes. In Body Sensor Networks (BSN) Workshop, pages 224– 229, 2009.
  5. D. Connaghan, P. Kelly, et al., “Multi-sensor classification of tennis strokes,” IEEE Sensors, pp. 1437–1440, 2011.
  6. R. Srivastava, A. Patwari, et al., “Efficient Characterization of Tennis Shots and Game Analysis using Wearable Sensors Data”, IEEE Sensors, pp. 1–4, 2015.
  7. http://www.st.com/content/ccc/resource/technical/document/datasheet/ef/92/76/6d/bb/c2/4f/f7/DM00037051.pdf/files/DM00037051.pdf/jcr:content/translations/en.DM00037051.pdf

Borrador Anteproxecto

Borrador

Idea posible sistema

  1. IMU Data Collection (Python on Android) → Using BLE (Bluetooth Low Energy), you collect IMU data on your mobile device.
  2. Send Data to Cloud → Use Firebase, AWS, or a REST API to send data to a cloud database.
  3. Process Data in the Cloud → Apply ML models or signal processing (on Google Cloud Functions, AWS Lambda, or a Python server).
  4. Store Processed Data → Save results in a database (Firebase, PostgreSQL, etc.).
  5. Retrieve Data in an Android App → The app fetches processed data from the cloud.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published