Published July 3, 2026 | Version v3

Wearable Sensor Data Fusion Training Load Recognition and Competitive Performance Prediction for Track and Field Training

Authors/Creators

Description

InterventionNodeMemory

A wearable sensor data fusion framework for training load recognition and competitive performance prediction in track and field training.

Overview

InterventionNodeMemory is a deep learning framework designed to integrate multimodal wearable sensor data, intervention variables, relational reasoning, and memory-based temporal stabilization. The goal is to recognize athlete training load and predict competitive performance more accurately and interpretably.

The framework is inspired by research on wearable sensor data fusion for track and field training, combining causal inference, graph-based relational modeling, and memory-augmented prediction. :contentReference[oaicite:0]{index=0}

Key Features

- Multimodal wearable sensor data fusion
- Training load recognition
- Competitive performance prediction
- Intervention-guided causal reasoning
- Graph-based sensor interaction modeling
- Memory-augmented temporal stability
- Robustness to missing or noisy sensor signals

Architecture

The framework consists of three main modules:

1. Intervention Guided Predictor
   Models the effects of training interventions such as intensity, environment, and athlete-specific factors.

2. Node Interaction Projector
   Uses graph-based relational modeling to capture dependencies among wearable sensor modalities.

3. Memory Stabilization Adapter
   Maintains historical representations to improve prediction stability over time.

Applications

- Personalized training optimization
- Athlete monitoring
- Fatigue and recovery analysis
- Sports performance forecasting
- Evidence-based coaching decision support

Experimental Results

The proposed framework shows improved performance in training load recognition, competitive performance prediction, robustness under sensor missingness/noise, causal intervention effect estimation, and long-term fatigue forecasting.

Repository Structure


InterventionNodeMemory/
├── data/
├── models/
├── configs/
├── scripts/
├── experiments/
├── utils/
├── requirements.txt
└── README.md
Installation
git clone https://github.com/your-username/InterventionNodeMemory.git
cd InterventionNodeMemory
pip install -r requirements.txt
Usage
python scripts/train.py --config configs/default.yaml
Citation
If you use this project, please cite the related work:

@article{burenbatu2026wearable,
  title={Wearable Sensor Data Fusion Training Load Recognition and Competitive Performance Prediction for Track and Field Training},
  author={Burenbatu and Ai Yisi},
  year={2026}
}

Files

Files (5.1 kB)

Name Size
md5:a1282541db309b29f8306acb0453e73c
5.1 kB Download