In this busy world, everyone wishes to have a personal robot to help them, especially bedridden
patients who need a helper all the time to take care of their necessities. Society seems to see
them as a burden. Therefore, we built a project, a Personal Assistant Robot for such bedridden
people that fulfills their very basic needs. Hence, there wont be any need of a human caretaker.
The main modules of this robot are: Speech Recognition, Object Identification and Vehicle
Movement via Raspberry Pi 3. For Speech Recognition, Dynamic Time Warping and for Object
Identification, Nearest Neighbor Algorithm is used. Our dataset includes 50 audio and 50 image
samples for each of the four words i.e. Water, Apple, Medicine and First-aid. Raspberry Pi 3 is
used to make it a standalone four-wheel assembly.
Personal assistant robot listens the voice commands via Bluetooth mic. The commands include
the names of the things to be fetched e.g. medicine, water etc. Speech Recognition System
running on Raspberry Pi 3 will be able to recognize the spoken word. The whole room will be
scanned via camera, rotated by stepper motor, to locate the demanded object. Once identified,
the robot will move towards the required object by sensing the distance via Ultrasonic sensor
and return to the patient with the object.