To design an android application to detect the aggression level, with importance to patients suffering from Alzheimer’s and dementia after acquiring the specific characteristics of the patient. Different parameters such as Heart rate, Facial Recognition, Blood Pressure, Motion Detection, Speech Recognition technology and most importantly EEG Signals.
Our approach to the problem was based on parameters such as blood pressures, heartbeat, motion detection and Speech recognition. Its implementation is done through a single android wearable, which will detect the different the parameters such as heart beat, blood pressure with a common sensor, whereas the other parameters like motion detection are done through accelerometer and speech recognition is done through a microphone which can either made be interfaced with wearable sensor and or be embedded in a patient’s cloth and then can be connected through a blue-tooth.
Our work is based upon the use of PHP, java, XML, we are also planning to make use of Google APIs provide for processing part and the UI output will be on Android platform. The problem that we would be facing is knowing the pre-determined values of the parameters, as a small change in it will produce unnecessary signals to the care-taker. Depending upon the level of these parameters above their base line, the android wear with the care taker will give corresponding signals so that the patient does not turn hostile.
According to the above displayed flow graph, the patient is monitored for the different values through the android wear and then the same is sent to the server, where the values are stored dynamically. These values are then converted to independent variables in the detection algorithm and the same is compared with their baseline values. We are taking the primary detection parameters as heartbeat, blood pressure and movement detection, whereas, the secondary parameter is Speech analysis, but with higher preference given.
The algorithm is such that, the first stage of triggering the text is based upon the values of heart rate, blood pressure, motion detection, if these values show abnormality, then the speech detection is checked for variations, if both the conditions are satisfied, then the next level is about the level of aggression detection, as in whether the patient is under normal circumstances, person is on the verge of aggression, the patient has turned hostile.
The webpage or the UI interface in the Desktop will display, continuously, the recorded values, ready to be accessed at any time. The android wear with the Nurse will also display these details, but in a more constrained way, such that it will be more concentrated in displaying a three-signal output, depending upon the aggression level, with taking into consideration a considerable amount of time for the care-taker to respond to the patients.
The processing system will have a computational unit, backend and frontend application. The frontend will display the output to the caretaker and also in the website, whereas the triggering part is done in computational part. Initially, the wearable watch detects the changes in physical parameters such as heart rate, blood pressure and sudden body movements of the patient. At the same time, the MIC will sense the changes in the decibel levels of the patient which may sometimes be the initial symptoms of aggressive behavior.
In the above diagrammatic representation, we can see that all these parameters that have been detected is sent to the processing unit to generate signals/data that will trigger the display. We have an existing database which contains the medical records of the patient contributing to the patient’s information. Using this, we can compare the obtained data from the sensors to the data which already exists. This will indicate whether or not the patient suffering from Alzheimer’s or Dementia is on the verge of showing aggressive behavior. The processing unit consists of a computational device which has backend and frontend applications.
It enables the data obtained from the server to be supplied to the website or an app that the caretaker is handling. The UI interface in the desktop will receive and display the obtained results, enabling the user to interact with the webpage. The android app that the caretaker will be using will also display the results in the form of three colors, green, yellow and red depending on the level of aggression. Each colors is assigned a specific range which will indicate different levels of aggression.