Abstract (EN):
Research on autonomous vehicles has come a long way since first findings, and its software tools are increasingly acclaimed by the research community. Particularly with robotics simulators, autonomous vehicles have been provided with a suitable test-bed for experimentation of new methodologies such as long-term navigation algorithms, map building and intelligent reasoning. However, when it concerns the deployment and validation of such vehicles in a larger urban traffic scenario, robotics simulators do not seem to provide the required functionality for road traffic analysis, or inter-vehicular communication infrastructure as they seem present in today's traffic simulators. The improvement of such features is the key for the successful practical deployment of such a critical system. In this paper, we present a study on the currently available robotics and traffic simulators focused in its applicability on autonomous vehicles research. The simulator integration approach was introduced by Figueiredo et al. (2009), although the selected software architecture and frameworks did not provide satisfactory results. In the former paper, the traffic and robotics simulation softwares chosen did not have the sufficient maturity level, which led to various implementation difficulties and, consequently, to an inefficient platform. However, its findings have provided some lights on the usefulness of newer possible integrations and difficulties to be overcome. In this paper we present the current features and issues on some of the most known robotics and traffic simulators, and propose a taxonomy for selecting these simulators, in order to foster further developments on the topic.
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
8