A major problem for many caregivers is ensuring that patients who are physically weak stay in bed. Our system attempts to solve this problem and consists of three major subcomponents. The first component is the single-board computer that will handle all of the local processing of the data received. In our project, we use a Raspberry Pi 4 running a distribution of Linux. The Raspberry Pi is used to interpret data sent from both our thermal and standard cameras to determine if a patient is in bed or not. This subsystem will also act as the bridge that interconnects the next two major components of this system: the first being a desktop application connecting through a local area network via Wi-Fi; and the second being a custom printed circuit board connected via a serial connection.

The desktop application previously mentioned will be used by healthcare providers to monitor patients in bed. This will be running in a central location such as a nurse’s station where it can be frequently checked for updates. The application will display the status of the patient in three different ways: “Safe”, “Warning” or “Alert”. The “Safe” state means the patient is okay and no further attendance is needed. “Warning” notifies the caregiver that the patient is exhibiting behavior that may lead them to attempting or inadvertently falling out of their bed. An “Alert” state means that the patient has fallen or is out of bed, and the alarm will sound. The application will sound an alarm and the patient identification whose alarm is going off will be prominently displayed. Another feature that is provided in the desktop application is the streaming functionality where caregivers can observe the patient without disturbing them.

Finally, there is the custom printed circuit board which is used to achieve one of the most challenging goals of this project. This component accurately detects the motion and state of the patient in the bed using an array of ultrasonic sensors and a microcontroller. When the board detects that the user is no longer in the bed, or that a user is attempting to leave the bed it will send a notification over a serial connection to the Raspberry Pi which will relay the information to the desktop application. If either the Raspberry Pi or the microcontroller detects a patient is out of bed using their respective sensors, they will relay the information to the other component and the microcontroller will sound an onboard alarm.