This application runs on an Android device.
Currently, the mobile version keeps the screen on and reads voice inputs using a custom dictionary. This is all thanks to the hard work from the PocketSphinx team at CMU (of which this repo is forked).
It has been modified to incorporate ROS. The app now processes voice inputs and Publishes the responses to a node on a ROS Master running on a different computer. We will be using this to communicate to our multiple drones. Additionally, the app also updates the view with information from a different ROS node that it is Subscribed to. Both these Subscribers and Publishers can be modified to support multiple nodes to Subscribe/Publish to.
Currently, the app only connects to 1 specific IP address, but that is set to change once we need it. (Needs to be voice togglable to tell which drone to connect to)
The wear app has not been updated with these changes, but can easy incorporate these changes if needed.