utexas-bwi / bwi_speech

Speech functionality for BWI

bwi_speech

Speech functionality for BWI

Installation and Setup

To install first clone this repo to your catkin workspace:

$ git clone https://github.com/utexas-bwi/bwi_speech.git

Now download all files pertaining to the submodules of the repo:

$ cd catkin_ws/src/bwi_speech/
$ git submodule update --init --recursive

Now that it's downloaded, switch to the speech acquisition branch:

$ git checkout speech_language_acquisition

Next, you'll need to install the other bwi packages following the instructions found here: https://github.com/utexas-bwi/bwi

Once that's installed, switch to the static fact query branch of bwi_common:

$ cd catkin_ws/src/bwi_common/
$ git checkout jesse/static_fact_query

Finally, you'll have to install any dependencies which the Google Speech API requires, this step may take some cleaning up of packages:

$ cd catkin_ws/src/bwi_speech/ThirdParty/Samples/Google/python-docs-samples/speech/grpc/
$ sudo pip install -r requirements.txt

Before you can use the Google Speech API, you need to install the Google Cloud SDK using the instructions found at: https://cloud.google.com/sdk/downloads#apt-get

After installing the SDK, you'll need to authenticate yourself to be able to call the API:

$ gcloud beta auth application-default login

Usage

In order to use, first source your workspace:

$ source catkin_ws/devel/setup.bash

Then open a new terminal tab to run roscore:

$ roscore

Then in another tab either run the Segbot simulator:

$ roslaunch bwi_launch simulation.launch

OR run the actual Segbot package for controlling one of the Segbots:

$ roslaunch bwi_launch segbot_v2.launch
$ roslaunch bwi_kr_execution bwi_kr_execution.launch

Finally, run the speech pipeline script:

$ rosrun bwi_speech speech_pipeline.py