Pheeni, the Stroke Rehabilitation Robot
Capstone
Past Work
The above poster was created by a past group of electrical engineers, who worked with Dr. Pierre Mourad to create a tabletop robot used in upper body motor recovery exercises for stroke patients. What sets this apart from current exercises, is that our robot is designed to add an element of 'fun' and a 'cute factor' to gamify stroke recovery, and combat patient burnout and depression. The past team was able to create a functional prototype, however had mechanical issues, such as a weak shell, poor wheel design, and inconsistent power delivery to wheels
Where I come in
I am currently leading a team of 4 mechanical and 1 electrical engineers, handling communication, planning, and project management. The project is very much in its infancy, however I have coordinated a project layout which should get us results fairly soon. The project will involve learning Arduino, motor selection, impact testing in ANSYS, patient testing at John Hopkins, mechatronics, sensor integration, image processing, rapid prototyping and PCB design. Progress will be updated below
IR sensors for line following functionality, protected by the upper shell of the housing
short range IR sensor to detect a patients reach, housed inside the upper shell of housing
Wheels housed inside upper shell of housing to protect drive train from drops
Front bumper protects lower IR sensors in case of drops, will likely require more reinforcement
Ball transfer bearing for stability
Arduino with motor shield pushed towards back to focus weight on wheels
9V battery, will later be replaceable through a hatch on the bottom, to allow continuous use
DC motors, may require controller for gradually applied torque to avoid slipping
Concept design
A basic model created in solidworks, to show some of the ideas we have spent the past week discussing. Only the battery, Arduino, and motor shield have accurate dimensions. Dimensions will be added for the motors and sensors once they are chosen, space for wiring, OLED and speaker will be added as well, and the model will be redesigned with more accuracy, for drop tests in ANSYS.
Game design
The purpose of our device is to eventually create a game which patients will find engaging, and challenging in order to prompt them to keep up with their recovery exercises. This block diagram I designed shows how the game will eventually work, where progress will be dependent upon patient performance, and may employ a 'lives' system where patients have must complete their assigned workout with a limited number of failures determined by the user/physician. During this all the robot will store information on distance from patient, time between touches, and successful interactions, which will create a 'score' for each session. This will all hopefully take place inside an Arduino board, which we have begun using to code our line following capabilities, using an array of 5 analog IR sensors and a PD controller.
Holloman Health Challenge
We were one of 16 groups selected from a pool of 51 applicants to participate in the Foster School of Business Holloman Health Challenge, a challenge which recognizes innovative ideas in healthcare, and seeks to invest $15,000 in the winners idea. We applied with an 8 page business summary, which they used to screen applicants. We then condensed that into a 1 page summary, and will be presenting a 60 second pitch in front of judges, and run a small demonstration afterwards.
​
For the demonstration, I designed the above model in Solidworks to show what our product may look like. Additionally, I wrote code in Arduino to control animations on an OLED screen, buzzer noises, and a flashing LED which all fit inside the 3D printed model.
Prototype progress
After I modeled and printed a base which could could hold the motors, IR sensors, and associated electronics (left), we were able to test some of the driving code we had written. The device is capable of following a path with only slight deviations, which will be factored out by implementing a PID controller. The code for which is shown below. Values on the PID controller can be changed in real time by using a potentiometer added to our Arduino circuit, to allow for testing on future frames. The PID controller unfortunately could not be implemented before our boost converter failed, which we have not been able to replace due to COVID-19.
​
An example of simple feedback using two i2c controller OLED screens is also shown (right). This was used as a prop at the Holloman Health and Innovation Challenge, to demonstrate the final size and possible look of our device upon completion. I designed and printed this prototype to mount two OLED screens, hold a button on top, and hold an Arduino nano and buzzer on a PCB board on the bottom to display our images.
PID controller method