A robust speech biometric system for vehicle access

Gaining key-less access to one's own vehicle using biometrics is a feature that is gaining popularity and has been implemented in concept and high level entry cars. But the state of the art in speech biometric especially as an embedded solution and the environmental conditions that the owner of the vehicle might be in (highway, noisy street, and garage) present challenges to the speech based verification system. We propose a multi authentication speaker verification system to enable the owner of the vehicle gain access to his vehicle. The idea is to have a small footprint embedded speaker verification system installed inside the vehicle and a robust speaker verification system (more accurate) installed on a server. The output of both the channels is fused to verify the user robustly.