Low Power Wide Area Network Architecture for Detection of Traffic Signal and Pedestrian Crossing in an Autonomous Vehicle

Abstract

Automotive Industry is rapidly evolving and the next revolution is the autonomous vehicles. These vehicles
will make mobility of human and goods safer and convenient. Autonomous vehicles sense the environment through
various sensors and process the data for decision making. In this study the various surround sensors are studied in detail
along with its limitations. As these vehicles need to meet the safety standard ISO 26262, this study also focused on the
various external threats to the network devices, sensors and the associated software.
In a road architecture there are two standard system available in a cross junction i.e. Traffic signal and pedestrian
crossing. In this study a new networking technology i.e., Low Power Wide Area Network is proposed for autonomous
vehicles which can efficiently predict the traffic signal and pedestrian crossing without any sensor embedded inside the
vehicle. This technology can cover wide area in orders of kilometers and also uses very less power compared to
available networking systems. This network uses an end device fitted in the signal and adjacent to pedestrian crossing
sending signals to the gateway installed in the vehicle and further connected the end server. Experiment are carried out
using LoRaWAN™ protocol in a virtual test bed environment to establish a connection between the end device and a
gateway. The results prove that implementation of LPWAN technology is a safer solution than utilizing complex and
costly surround sensor based system for detecting traffic signal and pedestrian crossing.