I am currently working with a Pepper robot (academic version and the QiSDK and NaoQi 2.9). Since I am using the academic version I can't use the cloud based automatic speech recognition service from Softbank which is not included and therefore e.g. I can't use wildcards or other chatbot engines besides QiChat. Does anybody of you know how I can implement my own speech recognition service for Pepper? I can't find where I can get access to the audiostream of Pepper's microphones.
I've read the documentation from Softbank: https://developer.softbankrobotics.com/pepper-qisdk and https://qisdk.softbankrobotics.com/sdk/doc/pepper-sdk/ch4_api/conversation/reference/basechatbot.html
And I've tried to create a SpeechRecognizer based on Android, which works, but uses the Tablets microphone and not Peppers.