Design and implementation of a system on chip for navigating and recognizing transit signals on a mobile robotic system

Abstract

Due to the fact that the traffic accidents are one of the most relevant causes of deaths in the world, this article aims to expose the design of a System on a Chip (SoC) that has the capacity to recognize some transit signals, obstacles and white lines that delimit a road of a controlled scenery, giving a future prospect in order to implement automatic  pilots  in  real  automobiles  for  helping  reduce  accident  rates.  The  SoC  operates  on  an  FPGA  (Field Programmable  Gate  Array),  so  that  the  high  speed  processing  algorithms  are  implemented  in  hardware  and  the  low  speed  processing  algorithms  are  implemented  in  software.  Therefore  different  hardware  modules  were  developed  for  video  capturing,  pre-processing  images  and  controlling  white  balance  in  video  capture.  Also,  software algorithms were designed, such as navigation, stereoscopy artificial vision and the intelligent recognition of the transit signals, shaping a set of processes administrated by a Real Time Operative System (RTOS). The SoC results were obtained through simulations in a robotic platform navigating on a road modeled on a small scale.