With the growing demand for infection control measures, device control by voice command is getting popular. While voice recognition such as in a smart speaker is popular among people, some manufacturers are recently considering applying it to embedded devices. The new approach is mostly considered based on its feasibility. Then, can you imagine how much processing you will need and how much recognition you can get from an embedded device?
We have released a demo based on our RX Family of general-purpose 32-bit MCUs, which can be installed into a commercially available kit and tried on immediately. If you have the kit, you can try it out and experience it yourself right now.
This time, we will introduce a demo using the RX72N, a flagship model of the RX Family. In this demo, the result of the voice command operation is fed back with graphics and voice playback. For example, when you give the command “Turn on the TV”, the TV on the screen is turned on with the voice, “I turned on the TV”.
To install this demo, we will need to first install the RX72N Envision Kit, which is available for about 60USD at some online shops. Next, download the demo code from the Renesas web page and write it into the kit using the free flash programmer software. For details about the demo, please refer to the attached Application Note.
Though the number of commands available is limited at this moment, you can still try voice recognition using an embedded device. You can do simple evaluation by adjusting the parameters such as the threshold values, noise suppression and beamforming.
After your confirmation of an acceptable result, the next step is to install the commands you would want to use and tune the hardware on the actual environment. At that time, please contact Renesas for further support or the vendor of voice recognition software.
We are looking forward to hearing from you!