Hyundai Mobis’s sound-based AI tech helps detect faults

[Courtesy of Hyundai Mobis Co.]

Hyundai Mobis Co. integrated a new artificial intelligence (AI) technology into its manufacturing process that detects defects based on the sounds that vehicle parts emit during operation.

According to the company on Wednesday, Hyundai Mobis has implemented an “Acoustic AI” inspection system at its Changwon plant on South Korea’s southeast coast, where it produces motor-driven power steering (MDPS) units.

Acoustic AI is a cutting-edge AI technology that interprets sounds using algorithms to recognize specific patterns and make accurate judgments, while MDPS is a device that uses an electric motor to assist steering. As the motor rotates inside an MDPS, it produces sounds with a consistent waveform. Acoustic AI analyzes this waveform to detect subtle variations, such as deviations or pitch differences, to ensure product quality.

Units that were suspected to be defective and below standard thresholds were previously identified post-assembly via automated systems and were manually inspected by specialists. With the new Acoustic AI system at Changwon, however, MDPS quality inspections can now be conducted at the rate of one unit per second, with Hyundai Mobis manufacturing 1.3 million MDPS units annually at the Changwon plant alone.

Acoustic AI’s key advantage is its capability to analyze quality issues and determine their root causes. The company developed this AI solution over the past year via collaborations with field engineers, production technology experts, and AI specialists. It intends to extend the application of this technology to other components that inherently produce noise during operation, such as braking systems.

“Acoustic AI is a significant innovation in process within the automotive sector and global manufacturing industry. We are committed to introducing advanced AI technologies for production, research, and development, and enhancing workplace environments,” the company said.