一、 Research background:
With the rapid development of intelligent robot technology, human-machine interaction (HMI) based on acoustic sensors plays a crucial role in promoting natural and efficient communication among robots. However, how to accurately identify and track omnidirectional sound sources, especially in noisy environments, remains an urgent problem to be solved.
二、 Article Introduction:
In response to the above issues, the team of researcher Wang Jie from Beijing Institute of Nano Energy and Systems has successfully developed a self powered frictional stereo sensor (SAS) with omnidirectional sound recognition and tracking capabilities, providing an innovative solution to solve this problem. SAS uses a porous vibrating membrane with high electron affinity and low Young's modulus, which gives it high sensitivity (3172.9mVppPa-1) and a wide frequency response range (100-20000 Hz). By utilizing its comprehensive sound recognition capability and adjustable resonant frequency characteristics, SAS can accurately recognize the desired audio signal even in noisy environments, with an average deep learning accuracy of about 98%. The development of this sensor not only solves the problem of sound recognition for intelligent robots in complex environments, but also opens up broad prospects for its application in multiple fields. For example, in the auxiliary conference system, SAS can simultaneously recognize the voices of multiple individuals, improving conference efficiency; In the field of autonomous vehicle, it can accurately identify driving commands under background music to ensure driving safety. These application displays signify significant progress in voice based human-machine interface systems. The relevant research results were published in Advanced Materials. The first author is Qiao Wenyan, Ph.D., Beijing Institute of Nanoenergy and Systems, Chinese Academy of Sciences, and the corresponding authors are Zhou Linglin, Associate Researcher and Wang Jie, Researcher, Beijing Institute of Nanoenergy and Systems, Chinese Academy of Sciences.
三、 Research content: 1. Sensor structure and working principle
Acoustic sensors play a crucial role in exploring more efficient and natural human-machine interaction (HMI) systems. As the "auditory" device of robots, acoustic sensors can accurately recognize human commands, speech content, and intonation, greatly promoting social interaction between robots and humans. This article proposes an innovative omnidirectional SAS that integrates five layered structures of self powered frictional electric sound sensors (TAS) on a 3D printed stereo frame, achieving omnidirectional capture and efficient recognition of sound signals. The working principle of TAS mainly includes two aspects: the deformation of FEP film caused by sound waves, and the vibration of the film actively converts the sound signal into an electrical signal.
By analyzing formulas (1), (2), and (3) and Figure 1, it is clear that the key indicator affecting TAS sensitivity is voltage (U); By adjusting the parameters of Young's modulus (E), radius (r), and film thickness (t), the vibration displacement of TAS can be changed, thereby changing the sensitivity of the voltage output regulating device of TAS. In addition, these three parameters can also adjust the resonant frequency (f0) of TAS. In order to achieve multi-directional sound recognition and real-time tracking of sound sources from noisy environments, the author introduced a 3D printed SAS design with five uniformly distributed surface cavities, each surface integrated with a single TAS. Based on omnidirectional sound recognition and tunable vibration frequency characteristics, SAS has demonstrated the ability to pick up target sounds in noisy environments, which has been demonstrated in autonomous HMI vehicles. To demonstrate the principle of SAS, the displacement responses of TAS and SAS under different sound source incidence conditions were simulated. When the sound source is facing TAS, the TAS has the maximum signal response. When the sound source is between two TASs, adjacent TASs have the same and maximum signal response. Based on these features, the orientation and angle of the sound source can be determined according to the response of SAS.
四、 Summary and outlook:
The author proposes a self powered SAS sensor that adopts a unique cube design, endowing it with dual capabilities of omnidirectional sound response and precise tracking. By combining a low-E porous vibrating membrane, SAS has high sensitivity (3172.9 mVpppPa-1) and a wide frequency response range (100-20000 Hz). By utilizing SAS's omnidirectional sound recognition and tracking capabilities, as well as its differentiated resonant frequency response to different sound sources and directions, efficient extraction of target signals from noisy backgrounds has been achieved. With the assistance of deep learning, SAS achieved an average recognition accuracy of about 98% for target signals. More importantly, the emergence of SAS has broken the limitations of multiple people interacting with robots simultaneously. In addition, SAS successfully demonstrated its outstanding performance in the auxiliary conference system, sound tracking and auto drive system (especially in the environment with background music to accurately identify driving commands). This study highlights the profound advantages of self powered frictional electricity technology in voice based human-machine interface systems.
Source: Sensor Expert Network. Reproduction of this article is to convey more information. If there is any infringement, please contact us for deletion