Voice command recognition on FLipMouse3
Supervisor: Benjamin Aigner aignerb@technikum-wien.at
Level: Internship / Bachelorproject
The upcoming version of the FLipMouse is based on the RP2040 CPU on an Arduino Nano RP2040 connect. This board includes an ESP32 (for WiFi) and a MEMS microphone
There could be different approaches for the implementation:
- Voice processing on the ESP32: If possible, the best solution would be a speech processing algorithm, which is directly processed by the ESP32 chip. This might be challenging, because the ESP32 is not a full desktop processor, so the usual speech processing frameworks are not suited.
- Voice processing on an external RaspberryPi: Another approach would be the outsourcing of the speech processor to an external board. For usability reasons, the FLipMouse and the Raspi should auto-discover themselves for an automatic handover of the speech data to the Raspi and the feedback of recognized commands to the ESP32.
Technologies / Requirements
C/C++/Python programming (the language depends on the selected speech processing framework), Linux, FreeRTOS