Multi-sensor fusion (MSF) is believed to be a promising tool for vehicular
localization in urban environments. Due to the differences in principles and
performance of various onboard vehicle sensors, MSF inevitably suffers from
heterogeneous sources and vulnerability to cyber-attacks. Therefore, an
essential requirement of MSF is the capability of providing a consumer-grade
solution that operates in real-time, is accurate, and immune to abnormal
conditions with guaranteed performance and quality of service for location-based
applications. In other words, an MSF algorithm depends heavily on data
synchronization, cost, an accurate process model, a prior knowledge of
covariance matrices, integrity assessments, and security against
cyber-attacks.
Multi-sensor Fusion-based Vehicle Localization addresses trending
technologies in MSF-based vehicle localization and outlines some insights into
the unsettled issues and their potential solutions. The discussions and outlook
are presented as a collection of key topics, including multi-sensor measurement
data processing, sensory selection, filtering, integrity assessment, and
cybersecurity.