An Intelligent Wheel-Chair Model using CNN-MPNN Based Text Classification

Individuals who are disabled by medical conditions like disability, injury, paralysis, or other circumstances struggle to move around. In most cases, assistive devices are needed, such as wheelchairs. Using voice instructions to control wheelchair movements is the main objective of this study. Wheelchair movements can be controlled by five commands: forward, reverse, right, left, and stop. By using the Google Voice virtual assistant, the voice can be captured, translated into English using the Google API, and then distributed to classification, which classifies the words and extracts the phrases. In-order to correctly identify the text commands for wheelchair navigation, Convolutional Neural Networks (CNN) and Multilayer Perception Neural Networks (MPNNs) were used and the implementation was performed using Arduino UNO board. The CNN-MPNN method has been shown effective in text classification. The CNN-MPNN can recognize speech commands in wheelchairs, according to preliminary testing findings.