Acquisition, processing and data analysis of piezoelectric sensors for training musical robots in a didactic model

Resumo

This paper introduces a developed system tailored for processing and analyzing data acquired from
piezoelectric sensors. The system’s objective is to detect rhythmic patterns in percussive music, generating a
database to train musical robots. This training enables collaborative interaction among musical robots, thereby
nurturing human musical advancement. The proposed approach involves assessing the musician’s performance
by installing piezoelectric cells affixed to rubber pads corresponding to each instrument. Processing techniques
applied to the piezoelectric sensor signals facilitate system implementation through accessible didactic hardware,

which is more user-friendly compared to alternatives such as frequency spectrum processing. Analyzing param-
eters based on rhythmic beat intensity and timing leads to the establishment of a precise database characterizing

musician dynamics. The progression from a simplified to a more intricate data model explores intensity and

data structure within the database. Testing employed well-known basic rhythms to construct a rhythm reposi-
tory, demonstrating the system’s adeptness in microcontroller-based data processing and analysis. The system

showcases benefits like compactness, energy efficiency, and reduced space and weight, favorable for constructing
robotic frameworks. The proposed approach supports real-time and embedded system applications, extending a
multitude of possibilities and applications within the realm of music and robotics.