Search Results

In this paper, a systems engineering approach is explored to evaluate the effect of design parameters that contribute to the performance of the embedded Automatic Speech Recognition (ASR) engine in a vehicle. This includes vehicle designs that influence the presence of environmental and HVAC noise, microphone placement strategy, seat position, and cabin material and geometry. Interactions can be analyzed between these factors and dominant influencers identified. Relationships can then be established between ASR engine performance and attribute performance metrics that quantify the link between the two. This helps aid proper target setting and hardware selection to meet the customer satisfaction goals for both teams.

The performance of a vehicle’s Automatic Speech Recognition (ASR) system is dependent on the signal to noise ratio (SNR) in the cabin at the time a user voices their command. HVAC noise and environmental noise in particular (like road and wind noise), provide high amplitudes of broadband frequency content that lower the SNR within the vehicle cabin, and work to mask the user’s speech. Managing this noise is a vital key to building a vehicle that meets the customer’s expectations for ASR performance. However, a speech recognition engineer is not likely to be the same person responsible for designing the tires, suspension, air ducts and vents, sound package and exterior body shape that define the amount of noise present in the cabin. If objective relationships are drawn between the vehicle level performance of the ASR system, and the vehicle or system level performance of the individual noise, vibration and harshness (NVH) attributes, a partnership between the groups is brokered.

Transient automotive sounds often possess a complex internal structure resulting from one or more impacts combined with mechanical and acoustic cavity resonances. This structure can be revealed by a new technique for obtaining translation-invariant scalograms from orthogonal discrete wavelet transforms. These scalograms are particularly well suited to the visualization of complex sound transients which span a wide dynamic range in time (ms to s) and frequency (∼100Hz to ∼10kHz). As examples, scalograms and spectrograms of door latch closing events from a variety of automotive platforms are discussed and compared in light of the subjective rankings of the sounds.

The work presented here is part of the research done in the field of voice biometrics. This paper helps to understand the state-of-the-art in speaker recognition technology potentially capable of solving challenges related to speaker identification (to identify a speaker among multiple speakers) and speaker verification/authentication (to recognize the current speaking person at a pre-defined access level and authenticate accordingly). The research was focused on performing an unbiased evaluation of two individual voice biometric services. The level of accuracy in identifying and authenticating individuals using these services provides an insight into the current state of technology and the state of what other dual authentication methods could be used to achieve a desired True Acceptance Rate (TAR) and False Acceptance Rates (FAR).

In this paper, an approach to analyze voice recognition data to understand how customers use voice recognition systems is explored. The analysis will help identify ASR failures and usability related issues that customers encounter while using the voice recognition system. This paper also examines the impact of these failures on the individual speech domains (media control, phone, navigation, etc.). Such information can be used to improve the current voice recognition system and direct the design of future systems. Infotainment system logs, audio recordings of the voice interactions, their transcriptions and CAN bus data were identified to be rich sources of data to analyze voice recognition usage. Infotainment logs help understand how the system interpreted or responded to customer commands and at what confidence level.