Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{201625,
        author = {Mrs. A. Vidhya and Dr.M.Rajaram and Mrs.C.Preethibha and S. Balakrishnan and S. Gayathri and R. VijayaLakshmi and Dr.G.Ganesh Kumar},
        title = {AI-Based Early Parkinson’s Detection Using Speech and EMG Signal Fusion},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {5552-5562},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=201625},
        abstract = {Parkinson’s disease (PD) is a serious condition with hidden and progressive symptoms, which is why its early diagnosis is a very difficult task. This paper introduces an artificial intelligence (AI) multimodal early Parkinsonian detection system that involves the combination of speech and electromyography (EMG) signals. Among the initial biomarkers of PD, speech impairments and neuromuscular abnormalities are suitable as they could be utilized in the non-invasive diagnosis of PD. The system proposed makes use of advanced speech signal processing methods that include Mel-Frequency Cepstral Coefficients (MFCC), pitch change, jitter, and shimmer analysis to achieve vocal biomarkers. Meanwhile, the envelope detection and frequency-domain analysis of EMG signals are to be used to detect tremor characteristics that are normally related to Parkinsonian conditions. The fusion is done at the feature level between speech and EMG features by complementary features, and then finally dimensionality reduction is done to enhance the efficiency of classification. Model compression and quantization-aware training and optimization of machine learning models to run on embedded devices allow real-timing machine learning inference on resource-constrained edge devices. It has been shown that multimodal fusion has a better detection accuracy, by experimental means, than when single-modality is used. The proposed structure places emphasis on the efficient combination of digital signal processing, embedded artificial intelligence and biomedical engineering concepts by introducing a cost-effective and scalable solution to the early screening of Parkinson disease.},
        keywords = {Parkinson’s Disease, Speech Signal Processing, Electromyography (EMG), Multimodal Feature Fusion, Digital Signal Processing, Embedded Artificial Intelligence, MFCC, Tremor Analysis, Machine Learning, Edge Computing.},
        month = {May},
        }