Pepper Speech Recognition

This is a python module that brings Google Speech Recognition to the Pepper robot by Aldebaran. It was specifically implemented so it can be run ON the robot as a NaoQi module but it can also be run on your computer.

Features

• Automatic detection if and when a person speaks
• Automatic detection when a person stops speaking
• Calibration function to learn background noise level
• Manual trigger mode (record now for 10 seconds and then recognize)
• This module provides an ALMemory event other modules can subscribe to
• Unlike other modules this one does NOT depend on PyAudio/Portaudio
• Only dependencies: Numpy, NaoQi Python SDK

Setup to run on robot

Everything you need should already be installed, simply clone the repository / copy the files to your robot, ssh into it and run

python module_speechrecognition.py

Setup to run on computer

Like every NaoQi Python project you need the NaoQi Python SDK. You can get it from doc.aldebaran.com. To test it simply run

python
import naoqi

Next up you need the numpy python module, which is the only python library used

pip install numpy

Test it:

python
import numpy

Now run the speechrecognition module

python module_speechrecognition.py --pip (your robot IP)

Configuration

To configure the module you can use an ALProxy instance you can get from the Broker (speechrecognition module needs to be running)

Here's a simple example on how to do that:

SpeechRecognition = ALProxy("SpeechRecognition")
SpeechRecognition.start()
SpeechRecognition.setLanguage("de-de")
SpeechRecognition.calibrate()
SpeechRecognition.enableAutoDetection()

Calling start() makes the module subscribe to the ALAudio event of the robot. Calibrate() makes it calculate the mean RMS value of the signal for 4 seconds and use this as a reference for background noise level.

Also, you can just use the module like this to trigger a one-time recognition:

SpeechRecognition = ALProxy("SpeechRecognition")
SpeechRecognition.start()
SpeechRecognition.setLanguage("de-de")

# starts immediately, records for at least HOLD_TIME seconds and then tries to recognize
SpeechRecognition.startRecoring() 

You can use module_receiver.py as a template for implementing your own module.

Subscribing to the ALMemory event

The speechrecognition module raises an ALMemory event every time a string was successfully recognized. You can subscribe to it in your module using the ALMemory proxy.

memory = naoqi.ALProxy("ALMemory")
memory.subscribeToEvent("SpeechRecognition", self.getName(), "processRemote")

"SpeechRecognition" is the name of the event, "proccessRemote" the name of your callback function.

def processRemote(self, signalName, message):
    # Do something with the received speech recognition result
    print(message)

ROS Module

A simple ROS (Robot Operating System) module is provided, that uses rospy to publish a ROS message whenever speech is recognized, carrying the recognized string as data.

Dependencies

The speechrecognition module was built to be able to run ON Pepper (and you can't easily install 3rd party libraries there, also we need to be efficient), so it's only dependencies are

• Python 2.7
• numpy
• naoqi Python SDK

All of them are pre-installed on Pepper

Built With

• Python 2.7
• PyCharm - Python IDE

Changelog

Version 1.0 - Initial release, May 30 2018

Authors

• Johannes Bramauer @ Vienna University of Technology - Initial work

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

• Anthony Zang (Uberi) and his SpeechRecognition