Development of Frameworks for Environment Dependent Traffic Simulation and ADAS Algorithm Testing
Author | : Shanthan Kumar Padisala |
Publisher | : |
Total Pages | : 0 |
Release | : 2021 |
Genre | : Automated vehicles |
ISBN | : |
Download Development of Frameworks for Environment Dependent Traffic Simulation and ADAS Algorithm Testing Book in PDF, Epub and Kindle
With the integration of Advanced Driver Assistance Systems (ADAS) and Intelligent Transportation Systems into vehicles, the need to measure the performance of these systems from a large scale traffic system level to vehicle component level is necessary in order to ensure the safety of the driver and all the traffic elements like pedestrians, other vehicles and infrastructure. Due to the practical constraints, software-in-loop (SiL) is the widely adopted methodology over on-road testing for verification and validation of these systems. However, these SiL solutions can be expensive and limited in their capabilities due to their proprietary nature. The unavailability of an open-sourced toolset for simulating microscopic vehicles at a macroscopic level has motivated the creation of a novel Simulation of Urban MObility based framework which can be used as a platform for system integration and co-simulations with other tools. Another problem addressed in this thesis is in the rapidly developing area of Perception System Algorithms. Due to the increased availability of data, these algorithms are being trained on huge datasets. However, due to the unavailability of proper evaluation methods or limited traditional metrics, it is challenging to evaluate the variation in the performance of these algorithms on images subjected to environmental variations. In order to evaluate the variation of the performance of an algorithm in different lighting conditions, a novel sensitivity based approach is proposed in this thesis.