At the 2024 Sensors Converge Conference and Expo, Fierce Electronics and Sensors Converge jointly presented the Best of Sensors Awards to honor outstanding advancements in sensor technology. The awards celebrate innovation, market value, and tackling critical challenges in the industry. This year, AiZip and Bosch worked together on their entry for the awards, the ZenVoice Bone deep noise reduction algorithm, and jointly won the award in the Medical and Wearables category.
The ZenVoice Bone algorithm, a joint development by Aizip and Bosch Sensortec, marks a significant advancement in noise reduction for true wireless stereo (TWS) earbuds. Their solution integrates Bosch Sensortec's BMA550 advanced bone conduction sensor with AiZip's sophisticated neural network design to achieve exceptional voice clarity in challenging acoustic settings.
Central to this technology is Bosch Sensortec's high-bandwidth accelerometer, which is specifically engineered for bone conduction applications. The sensor excels at capturing voice vibrations transmitted through bone, enabling precise filtering of ambient noise. Its performance surpasses that of conventional accelerometers, providing clear voice signals necessary for effective noise cancellation and speech enhancement. The sensor's energy efficiency and broad frequency response make it suitable for continuous voice detection in compact wearable devices.
AiZip's contribution is its neural network architecture, which complements the hardware with highly efficient audio processing. Their models significantly reduced computational demands and memory usage without sacrificing noise-reduction capabilities. This efficiency allows for the implementation of complex audio processing on small, power-constrained chips, making it ideal for integration into wearable devices.
The ZenVoice Bone system can preserve speech intelligibility while significantly reducing background noise, even in extremely challenging conditions where ambient noise exceeds the volume of the speaker's voice. It can also isolate the user's voice from surrounding conversations without prior voice training, a capability not available in traditional single-microphone systems.
The algorithm operates smoothly on hardware with limited processing power, requiring minimal memory. This compact footprint allows for real-time noise cancellation with imperceptible delay, making it useful for a wide range of audio devices beyond just earbuds, including various types of hearing enhancement products.
This collaboration between Aizip and Bosch Sensortec illustrates the synergistic relationship between advanced sensor technology and sophisticated AI algorithms. The high-quality data captured by Bosch's sensor provides the foundation for Aizip's algorithm to perform its intricate noise suppression and voice enhancement.
This integrated hardware-software solution pushes the boundaries of audio technology by addressing the persistent challenge of clear communications in noisy environments. Its potential applications extend beyond consumer electronics to professional communications, healthcare, and industrial safety where clear voice transmission is essential.
Aizip, a Silicon Valley-based company founded in 2020, has quickly established itself as a leader in model design for AIoT. Their proprietary automated design tools and neural network architecture have enabled them to develop high-performance deep neural networks (DNN) for a wide range of applications.
Bosch Sensortec is known for its innovative MEMs-based sensing solution in consumer electronics. A subsidiary of the Bosch Group, the company's portfolio includes a wide range of motion, environmental, and acoustic sensors that enable Innovative features in smartphones, wearables and other IoT devices.
To learn more about the other innovative companies and products recognized at the Best of Sensors 2024 Awards, visit the full list of winners here.