The are many types of robotic arms, but the main type for this project is a six axis arm driven by 6 servo motors. Each of them controls a specific motion in a region of the arm. The more axis found in a system the better, because there would be more degrees of freedom.
In order to control each servo, a pulse with duration t must be sent to achieve desired angle. It is not very straightforward to begin controlling these types of motors, but once the right timing is accomplished it becomes easier to drive them. Look at the data sheet for your specific device.
The process where a computer is able to distinguish between different sets of data without being programmed to do so is called machine learning. The most valuable feature of machine learning for the purpose of this project is voice and image recognition. For the demo below I designed a custom BoosterPack for the MSP432 to control the BRACCIO because currently there is no compatible hardware for the MSP432 and the BRACCIO that can be mounted as a BoosterPack. I am using a USB webcam for the object recognition.
The addition of an android application to control the BRACCIO can be useful to take advantage of the sensors found in phones. I was able to control the motion guided with my wrist movement because the phone was strapped to my wrist.
To achieve motion through voice and image recognitionFirst we need to train our model to do image and audio recognition. Then we perform a set of finite functions for the BRACCIO corresponding to the received data. In this case, we should choose a sequence that will allow for the BRACCIO to remain stable while performing a motion response.
In order to add an extended functionality I implemented my phone using the IMU unit that allows for my arm's motion to give a specific command. There are many ways one could implement this, but it must have a very reliable device with an IMU built in.
Valuable link resources (external):Getting started with machine learning
Comments
Please log in or sign up to comment.