Please login first
An FPGA Platform Proposal for real-time Acoustic Event Detection: Optimum platform implementation for audio recognition with time restrictions
* , *
1  GTM - Grup de Recerca en Tecnologies Mèdia, La Salle - Universitat Ramon Llull


Nowadays, monitoring of people and events is a common matter in the street, in the industry or at home, and acoustic event detection is commonly used. This increases the knowledge of what is happening in the soundscape, and this information encourages any monitoring system to take decisions depending on the measured events. Our research in this field includes, on one hand, smart city applications, which aim is to develop a low cost sensor network for real time noise mapping in the cities, and on the other hand, ambient assisted living applications through audio event recognition at home. This requires acoustic signal processing for event recognition, which is a challenging problem applying feature extraction techniques and machine learning methods. Furthermore, when the techniques come closer to implementation, a complete study of the most suitable platform is needed, taking into account computational complexity of the algorithms and commercial platforms price. In this work, the comparative study of several platforms serving to implement this sensing application is detailed. An FPGA platform is chosen as the optimum proposal considering the application requirements and taking into account time restrictions of the signal processing algorithms. Furthermore, we describe the first approach to the real-time implementation of the feature extraction algorithm on the chosen platform.

Keywords: acoustic sensor; signal processing; machine learning; real-time; FPGA; VHDL
Comments on this paper
Dirk Lehmhus
Miniaturization issues
Dear authors,

thanks for the interesting contribution. I have one rather generic question - as we are interested in material-integrated systems, how, do you think, could your platform be adapted to a strain sensing scenario, i. e. as a part of a sensor network with the sensor node being capable of signal and at least initial data processing? What could the minimum footprint (e. g. geometry, volume) of the system be if you'd make use of all that is available in terms of miniaturization approach while maintaining the claim to stay "low cost"?

Best regards,

Dirk Lehmhus
Marcos Hervás
Dear Dirk Lehmus

Thanks for your comment. The platform used in this First approach is the basys3 from digilent based on an Artix-7 of Xilinx (XC7A35T-1CPG236C). This platform is provided with 4 pmods connectors able to increase easily its capabilities of sensing, storing data, displaying results or communicating to a remote server. You could have a look over at to see an example of these peripherals.

This platform has a lot of peripheral we are not using, and the size could be smaller. In fact, the manufacturer Trenz-electronic have very small reduced modules such as the presented in the following link: The size of this module is 4x5cm and is provided with a high density connector allowing the user to design its sensing platform.

Best regards,
Marcos Hervás

Daisy Lucas
Best shows
Visiting such events in many ways helps you gain new knowledge and this is really an interesting topic because I understand this. Today you need to attach attention to many different events for the rest to at least relax a little. Therefore, I advise you to buy vegas show tickets the site with the best shows. This will cheer you up and give you new emotions.