Signal Recognition


This module aims to develop an adaptive model of inertial sensing, focused on vehicular perception and proprioception. Signals from a variety of non-intrusive sensors, embedded in modules attached to the vehicle, will be used to recognize and classify road surface patterns and driving events. Contextualized in a dynamic system permeated by uncertainties arising from noise and imprecision, the data captured will go through pre-processing to be developed, aiming the removal of deviations and wave smoothing, with application of signal filters and sensor fusion; besides the reorientation of the axes between the monitored coordinate systems, with the normalization of signals. After this processing, the resulting data will be modeled and applied in different techniques for pattern recognition, especially in expert systems, in order to address the current research gaps. The resulting classifications should be made available for asynchronous integration. The main contributions of this research include adaptability and portability of the model in relation to the context of experimentation, and the correlation of models through combinatorial study of the techniques for different patterns recognition.


Vehicle Perception Based on Inertial Sensing

The vehicle perception based on inertial sensing seeks to realize the recognition of the lane in which the vehicle travels, through confirmations by evidence from methods based on vibration. Contextualized in a dynamic system based on the laws of classical mechanics, acknowledgments made through the captured signals can be sub-categorized in identification and classification of pavement irregularities and road surface quality or composition.




Able to measure the linear acceleration in the direction of a reference axis. Acceleration is the rate of change of velocity in time, represented in m/s2 (meter per second squared) in the International System;

  • Measures applied inertial forces (acceleration), including the gravity component;
  • Velocity.


Capable of measuring the angular velocity around a reference axis. The angular velocity is a magnitude representing the rate of change of the angular position in time, whose unit of measure in the International System is rad/s (radian per second).

  • Directional reference.


  • Position reference. Angles in relation to the real world coordinate system;
  • Employee in the mapping of forces of the accelerometer.


  • Gravitational acceleration force on the three physical axes of the sensor.


  • Location;
  • Velocity.

About the Author

Thiago Rateke is a Computer Vision Researcher with experience mainly focusing on visual perception for autonomous navigation. Finished his PhD degree at Federal University of Santa Catarina (UFSC) in 2020 with focuses on visual perception for Autonomous Navigation. Using approaches like: Stereo Vision, Optical Flow and Convolutional Neural Networks.