End-of-line inspection for annoying noises in automobiles: Trends and perspectives

Annoying noises such as buzzes, squeaks and rattles (BSRs) are particularly important to the perception of quality in complex manufactured durable goods such as automobiles. These products represent significant expenditures to the average consumer and are expected to provide safe and reliable performance for many years. Consumers use sensory responses such as sound as a gauge of the product quality; hence it is important to provide a product that does not exhibit BSRs. Manufacturers have traditionally relied on human auditors to conduct product acceptance testing for BSRs at the end of the assembly line.This paper provides state-of-the art review and future perspectives for the current processes of BSR inspection used by automobile manufacturers in the United States. The requirements for BSR inspection methodologies are presented in terms of three functional areas: detection, classification and localization. While human auditors provide an effective combination of these functions, they are limited in their performance. A critical perspective analysis of the current process, in terms of these three areas, reveals significant issues with repeatability and reproducibility of inspection results when human auditors are used. These issues demonstrate a clear need for more objective methods of inspection for annoying noises at the end of the assembly line. A number of new technology trends, particularly in the fields of psychoacoustic applications and acoustic imaging hold some promise in addressing the limitations of human auditors. These are discussed in the context of requirements for an ideal inspection system which is capable of providing all three capabilities. There is research being conducted in each of the three functional areas, and in some cases at the intersection of two out of the three areas, further work is needed in developing a methodology combining these functionalities.

► Analysis of methods used at automotive assembly plants for finding annoying noises in vehicle cabins.
► Discussion of new technology trends that address the limitations of the current human-based methods.
► Acoustic beamforming has the most potential as a methodology to automate the inspection process.
► Spherical beamforming array tested shows sensitivity to source frequency.
► Further work recommended combining beamforming data, pattern recognition and psychoacoustics.