Critical Healthcare Assessment using WBAN and SVM

Abstract

Today good healthcare facilities and awareness of need of good healthcare is increasing in India. But as awareness increases it also strains the current healthcare infrastructure as patient expects more secured treatment round the clock. So there arises a need of remote assessment of patient health all the time using IoT devices. But these devices also need to be monitored by health worker in a hospital. Due to human interaction with theses IoT devices it may give rise to errors as human decisions can be late as a human health worker cannot look at the devices 24X7. So, to remove dependence of human decision-making technologies such as WBAN, cloud and machine learning has to be utilized together to make heath decision of a patient with less human interaction. So, we are designing a project where healthcare of a patient can be monitored extensively using WBAN. In first part of our project, we design a IoT device using Arduino and ESP8266 Wi-Fi module. The sensors connected to the Arduino will be pulse sensor, temperature sensor etc. The sensors will transfer data from patient to a server using ESP8266 and Wi-Fi called as WBAN network. The server will then apply SVM machine learning algorithm on the sensor readings and classify in two categories safe and unsafe. Custom made training dataset will be used to train the SVM. If unsafe readings are found the sensor will send a message to concerned doctor and upload readings to the cloud. The doctor on receiving alert can see the readings on the android app designed for the project and take a decision on the condition of the patient. For the project we are using Google Cloud Platform as our cloud provider which is free for use. Thus, by using our project a doctor can monitor his patient remotely from anywhere and the system will help in making decisions on the behalf of the doctor.