Smart India Hackathon Hardware Edition 2018
Humanoid Nurse
Humanoid Nurse
SIH Hardware was a whopping 5 days long event in which we barely slept and worked tirelessly to build our solution. Our solution was awarded the first prize.
During my time at SIH Hardware, I
Worked with complex embedded systems and prototyping tools such as multiconnected Arduino boards and Raspberry Pis
Build complex and neat circuit boards
Worked and implemented the complex math for translation using Mecanum wheels
Built an application for controlling the robot remotely
Worked with IMUs to detect body pose and transfer it using actuators on our bot
We were given the task of making the lives of people working in hospitals easier using hardware based technology. The specific problem that we identified a majority of nurses and supporting staff in the hospitals face is when they have to move/turn the patients around. This task requires atleast more than one person and is physically dangerous to not only the patient but also the people lifting him.
Our initial idea was to build a smart robot which could autonomously do all these tasks by itself. But that's more dangerous and harder to be accepted by the general market. Building on this opinion, we decided to combine the mechanical power of the robot with the actions of the nurse who has been trained to work delicately with patients.
Our solution consists of two systems:
A Vest equipped with motion sensors
A robot with the appropriate actuators
I am wearing the vest equipped with IMUs and the motion data is filtered and then sent to the robot and it mimics my actions
Using the motions detected from the vest worn by the nurse, we apply move the actuators of the robot so that it performs similar actions. By placing an inflatable tube on the side of the patient and using an air compressor, we facilitate the lifting.
For moving the robot around I used mecanum wheels, since they allow arbitrary translation without any rotation required. This is especially useful for smaller rooms with sensitive equipments that will limit the degrees of freedom for the robot. I also built an android application for controlling the robot maneuvers through bluetooth.
Other prototyped features included detection of the body pose and the detection of sensitive equipment such as catheters and wires, for making better weighted decision by the bot.
Our system can also be used for training purpose, wherein an expert wears the vest with motion sensors, and trainees wear a vest with actuators, and the motion of the trainee is subtly translated to mimic the motion of the expert. Other major applications of this system includes remote controlling heavy duty bots to lift off debris and rescue operations after disasters.