An AI assistant like Google Now, Siri, Cortana, etc. for an LTU senior project.
This project depends on the following:
espeakfor audio output- stanford_corenlp_pywrapper for natural language processing
- The Stanford CoreNLP jar files
- Java (to run Stanford CoreNLP)
- The Python SpeechRecognition module for voice input
- This might also require that the PyAudio packages are installed. On Ubuntu,
this can be done by running
sudo apt-get install python-pyaudio python3-pyaudio. - It also may require flac command line tools. On Ubuntu, this can
be done by running
sudo apt-get install flacAll of these are downloaded and installed by runningsetup.sh, as long as you have git installed.
- This might also require that the PyAudio packages are installed. On Ubuntu,
this can be done by running
It currently runs on Ubuntu-based Linux systems, though we would like to implement cross-platform functionality soon.
The main program is located in LTUAssistant.py. A string can be passed to it
on the command line as an argument if you want to execute a specific command
without speaking it.
This AI assistant aims to be like other well-known AI assistants, but with a focus on being useful to students. Thus, the feature set will concentrate on this particular target user.
Some abilities the AI assistant will include are:
- Ability to go to school websites via voice command
- Ability to email professors and classmates using LTU webmail
- Room locator for a given room number (i.e., building and floor)
- Study assistant using user-defined "note cards" to help students prepare for exams and quizzes
- Reminding the user of upcoming class times
This list may expand based on whatever we have time to implement.
This assistant is structured into three main parts:
- Speech-to-text module
- Natural language processing module
- Command execution and database module
Each part passes its results to the next, providing a logical program flow and making collaboration between our group members easier.