Date of Award

1-21-2016

Publication Type

Master Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

Wu, Jonathan

Keywords

Autonomous vehicles, Data Association, Multi Object Detection, Multi Sensor Data Fusion, Multi Target Tracking, State Estimation

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Abstract

Multi sensor Data Fusion for Advanced Driver Assistance Systems (ADAS) in Automotive industry has gained a lot of attention lately with the advent of self-driving vehicles and road traffic safety applications. In order to achieve an efficient ADAS, accurate scene object perception in the vicinity of sensor field-of-view (FOV) is vital. It is not only important to know where the objects are, but also the necessity is to predict the object’s behavior in future time space for avoiding the fatalities on the road. The major challenges in multi sensor data fusion (MSDF) arise due to sensor errors, multiple occluding targets and changing weather conditions. Thus, In this thesis to address some of the challenges a novel cooperative fusion architecture is proposed for road obstacle detection. Also, an architecture for multi target tracking is designed with robust track management. In order to evaluate the proposed tracker’s performance with different fusion paradigms, a discrete event simulation model is proposed. Experiments and evaluation of the above mentioned methods in real time and simulated data proves the robustness of the techniques considered for data fusion.

Share

COinS