Being a volunteer at the Alzheimer's Disease Association for over 2 years, I have dealt with numerous cases of elderly wandering off suddenly and losing their way. This worries the loved ones of the family a whole lot.
For this project, I am planning to build a smart iot clip that securely clips to the elderly clothing.
Requirements:
1) Accelerometer/motion sensor - With motion and acceleration sensor, I would be able to collect enough motion data and can detect when the elderly is lost using machine learning. (I would first create a training dataset by attaching it to the elderly for the first few days and get the baseline motion of the elderly. I would use an ARIMA model to model the time series data and would try out a few algorithms, such as recurrent neural network to detect anomalous data.
2) Microphone for Voice Recognition - With sound data collected from the microphone, it would allow again for words that can indicate that the elderly is lost. For instance, it can be activated with words such as "Are you lost" or "Where am I". This would add as an extra feature to allow us to detect whether the elderly is lost.
When there is high level of confidence that the elderly is lost, we can use sigfox network to send messages to their loved ones, or elderly care centre around. The MKR Fox 120 would be great for transmitting the message using the sigfox network.
Comments