Innovative Detection of Human Motion Intention using a Hybrid EEGNET Framework

The possibility for improved non-muscular control and communication has piqued the curiosity of researchers in human intention recognition using EEG data. In order to comprehend the relationships between brain activity and behaviour, recent developments have centred on feature extraction from EEG data. But using signal correlations among nearby EEG sensors is difficult, and typical feature representations based on vectorisation experience problems with high levels of signal noise. For accurate Human Motion Intention Detection, it is essential to have a standardised representation of frequencies, since various frequencies contribute differently to activity recognition. Data preparation, feature extraction, and training of models make up the suggested method. The Common Spatial Domain approach is mostly utilised for feature extraction, and artefact mitigation is achieved using linear filtering and avoidance. Following this, a hybrid EEGNET model is trained, which considerably surpasses the performance of traditional methods like CNN and LSTM networks. The usefulness of the proposed approach in Human Motion Intention Detection was demonstrated by its excellent success rate of 97.52%. This research provides a very effective method for detecting human motion intentions using EEG, which has important implications for the development of non-muscular control and communication systems.