Abstract:
To improve safe driving and prevent traffic injuries
caused by human factors such as speeding, fatigue and
distraction, techniques to understand the surroundings of the
vehicle are critical. We believe that being able to detect and track
every stationary object and every moving object, to reason about
the dynamic traffic scene, to detect and predict every critical
situation, and to warn and assist drivers in advance, is essential
to prevent these kinds of accidents. As an initial step towards
understanding whole scenario around a vehicle, in this paper we
have made an attempt to address the issues related to roadway
environment monitoring by only utilizing laser measurement
systems (LMS) as a perception sensor. Extensive experiments
were carried out to analyze the robustness of the proposed
methodologies in real campus and city environments. The results
are appealing and robust to various traffic and road scenarios.
