Ponentes
Descripción
In this presentation, we propose the application of the Brain Emotional Learning Based Intelligent Controller (BELBIC) as a bio-inspired intelligent filter for inertial sensor data. The work begins with a conceptual and functional analysis of the BELBIC model, which is rooted in the emotional learning mechanisms of the human limbic system. Its adaptive nature and ability to respond to stimuli without relying on a precise mathematical model make it an attractive alternative for filtering noisy data.
Subsequently, we introduce the integration of an Inertial Measurement Unit (IMU) and define a specific dynamic model for signal evaluation. The selected sensor for this study is the Bosch BNO055, which provides motion and orientation data. We present a comparative evaluation of three signal types: raw IMU outputs, the model-based output, and the output filtered through the BELBIC architecture.
Preliminary results indicate that BELBIC significantly improves signal quality by attenuating noise and preserving relevant dynamics, highlighting its potential as an adaptive and efficient filtering approach in real-time systems and sensor driven applications.
