January 2023 - March 2023
Building a Bionic Hand
I was part of a 10 week program where I was tasked with creating a bionic hand in 10 weeks alongside a group of 3 other members. The hand would have to complete 3 main tasks 1) stack blocks on top of each other, 2) sort shapes into designated holes, and 3) hold a grape. This project developed my skills in using arduinos and sensors, as this project introduced me to using flex sensors for the first time.
Ideation
Ideation .
Sketch 1
For the first sketch, a simple gripper was developed as the easiest solution to the three sketches. This would be actuated with a single servo, and gear mechanism. Pressure sensors would also be attached to the inside of the gripper so that a pressure-feedback loop can be created to handle delicate objects like grapes. Furthermore, as the entire assembly has minimal parts, it would be easy to iterate and manufacture. However, this is the least human-like mechanism, and is not the most effective at holding a wide array of objects and shapes.
For the task of our class, we were assigned the task of creating a bionic hand that would be able to complete 3 main tasks.
Task 1: Sort different shapes into appropriate slots
This task is the same as a game kids play when they first learn shape recognition and are improving their hand eye coordination. In this way, the robot’s ability to complete this task would involve good dexterity.
Task 2: Stack as many blocks as you can on top of each other
This task involves good hand eye coordination, further testing a robots dexterity, as this task now requires a precise placement of the block as well as the ability to place an object without disturbing others.
Task 3: Pick up grapes and place them into a bowl.
While this task again trains hand eye coordination, further testing the robot’s dexterity, it also involves gentle handling of the grape to prevent it from being squished, hence proving the robots ability to react to different types of objects with different pressures, as would be desired from a real bionic hand.
Before developing the hand, we sketched up 3 ideas.
Objectives
Sketch 2
For a more human-like bionic hand, we developed sketch 2, where we would use 2 servos. One servo would be attached to a strig that looped through the inside of the fingers, and would be actuated to close the fist. The second servo would be attached to a string on the outside of the finger, that would be actuated to open the fist. The hand would also be wrapped with grippy material to improve grippability. However, this sketch would only have an open and close fist capability, which is impractical when it comes to gripping smaller objects like a grape.
Sketch 3
For the final design, and the design that was used, sketch 2 was developed to actuate the thumb and forefinger seperately to the reaminder of the fingers. In this way, a fist gripping actuation could be created with actuating all the servos at the same time, while the forefinger and thumb would be actuated by their own independent servos to allow for the gripping of a grape. In the final design, the amount of servos was reduced by removing the need for a string to be threaded through the back of a finger by simply using elastic bands to snap the fingers back to their original position.
Design .
For the design of the bionic hand, a hand was taken from thinigverse, and adjusted to fit the servos we had available to us, and create holes to allow us to thread in string and attach our elastic bands. SolidWorks was used for the CAD design, and the wrist was designed independently. To house our electronics, a piece was laser cut and designed to slot in between the two halves of the palm. The CAD can be seen below.
Manufacturing.
Difficulties
The majority of the bionic hand was built using 3D printing technologies. FDM 3D printers were used to build the palm and coverrs for the palm, as well as the fingers and wrist. A forearm was also generated using lasercutting technologies. The finished product can be seen in the image to the right.
While manufacturing the hand, I experienced a good deal of difficulty with getting the finger to actuate in the manner I desired.
While manufacturing, I realised that I should have created a better place to mount the motors as the strings mounted on the palm as designed were too short to get easy motion with the motors. By using longer strings, the motors exerted a larger force allowing the finger to curl as desired (longer strings allowed for larger moments as they acted as longer levers). To rectify this, I designed and laser cut a piece of wood to connect to the wrist and act as a forearm, and this was where the motors would be mounted to.
Another issue I faced with the fingers was getting the fingers to spring back. From our design, we had decided to use some kind of elastic band to spring the fingers back to their resting state. However, with the design of the finger segments, it was rather difficultl to thread the elasticin a way that kept it untangled and get the finger segments to spring back. To fix this, I ended up drilling a hole through the bottom of some of the finger segments, and the top of the finger segments, so that the string could be more easily thread through. With more time, I would have made these changes in SolidWorks, and reprinted the design. After this, I used larger bolts than I originally intended to secure the strng behind the bolts that connected the finger segments together. Through this, the string retained tension better. The elastic was then threaded through the new holes created, so that when the string lost tension from the motors, the elastic that would have been stretched when in a curled position snapped back to its original state.
While testing with gripping objects I realised that the position of the fingers actually made it hard for the hand to hold onto the objects as the palm, rather than being flexible, was also perfectly flat. To fix this, I created a small bracket that could attach the thumb, and tested out different locations on the palm until the hand was able to decently grab objects. With more time, these features would be reflected in CAD and a reprinted design.
Electronics.
The bionic hand was actiated using 3 servos, and used a glove with flex sensors to provide input to the bionic hand. An arduino was used to power the electronics, where two functions were created. As explained earlier, the 3 servos were used to actuate the pointer finger, thumb, and the remaining fingers. This would be done by wrapping the strings that mechanically actuate the fingers around the servos and tying it, so that when the servos rotated, the strings would be pulled as well.
The first function was coded to actuated all 3 servos, creaing a gripping mechanism. The second function was coded to actuate 2 servos corresponding with the pointer finger and thumb. The wrist was attached to another motor to mimic an elbow, which was externally actuated using a controller.
To minimize the amount of flex sensors required, therefore, only 2 flex sensors were needed to trigger the 2 different functions. The degree to which the flex sensor was bent too corresponded with the degree the servos would rotate and pull the strings.
