Enhancing Transportation Safety with Infrastructure Cooperative Autonomous Driving System
|
Jinhong Noh , Yongwoo Jo , Jooyoung Kim , Kyongwook Min |
Superintelligence Creative Research Laboratory , Electronics and Telecommunications Research Institute (ETRI) |
|
|
|
|
ABSTRACT |
As self-driving cars become a reality, concerns about the safety guarantee of autonomous driving also increase. We propose
an autonomous driving system that enhances its vision capacity by leveraging edge infrastructure. This proposed system
involves autonomous vehicles participating in V2I broadcasting of the edge infrastructure, which supplies object perception
information. The vehicles then form policies from this acquired information. These policies, defi ned within a set of
constraints, enhance safety by impacting the trajectory of the autonomous vehicle. We implemented this system in both realworld
and simulated environments. An aggressive scenario at an unsignalized intersection was also performed to evaluate
the safety of the proposed system. The results showed that the edge infrastructure improved the safety speed by an average
of 17% and averted collisions with objects moving at speeds lower than 25 kph. Therefore, our strategy for enhancing safety
by expanding the fi eld of view for self-driving vehicles is successful. The proposed system is expected to be highly utilized
as it has an adaptable structure that can be easily expanded from existing autonomous driving systems and can consider
various types of traffi c participants. |
Key Words:
ICAD system · Autonomous driving · Cooperative perception · V2I communication
|
|