Socially Assistive Robot to Treat/Assist Alzheimers Patients

This study presents a Socially Assistive Robot (SAR) to assist Alzheimer's patients with emotional support, cognitive engagement, and real-time monitoring. Alzheimer's patients often battle with problems like memory loss, forgetfulness, and diminished social engagement, while carers deal with a tremendous burden and stress. The primary concern raised is the lack of affordable, continuous, and tailored monitoring devices for at-home dementia care. The proposed SAR integrates voice recognition, facial identification, emotion detection, and therapeutic interaction through memory games, music therapy, and safety alarms. The combination of autonomous response, patient classification, and real-time emotion-aware engagement on a lightweight Raspberry Pi platform makes this study unique. By classifying emotions and identifying patients using Deep Face, FER, and LBPH models, the system achieves 70% accuracy in emotion recognition and reliable speech understanding in indoor contexts. The experiments' results show effective patient identification, context-aware reactions, and successful cognitive involvement. The findings demonstrate that the developed SAR can improve patient well-being, reduce carer stress, and foster continued friendship.