Vibrations carry a wealth of useful physical information in various fields. For example, in nature, spiders can localize the struggling insects through identifying the transmitted vibrations. In health care fields, human health information can be obtained by monitoring heartbeats and pulses. In industry, the status of the industrial equipments can be estimated from vibration signals to ensure the operation safety. Sensing and locating multi-source vibrations has become an attractive research topic. It also has applications in other domains such as smart devices, robotics, and Internet of Things.  

Before being picked up by sensors, vibrations are filtered and mixed during propagation. The main approach to vibration source identification is to solve the inverse dynamical problems by measuring the related quantities. Traditional methods usually use distributed sensors or sensor networks to identify multi-source vibration information from the mixed vibration signals. These methods usually require complex data acquisition systems and control circuits, and results in high power consumption.  

Figure 1: Illustration of randomized resonant metamaterial for identifying vibration sources with a single sensor（Drawn by Guoyan Wang and Lei Chen）

In this work, we present a strategy of designing the randomized resonant metamaterial to identify multi-source vibrations with a single sensor. Here, metamaterials are artificially designed composite structures with novel properties not found in nature. Unlike traditional methods, the randomized resonant metamaterial can spatially encode vibration transmissions through careful design. We first develop a theoretical model with randomly coupled resonators. The disordered coupling of effective masses is the physical basis to produce the highly uncorrelated vibration transmissions.  

Figure 2: Metamaterial model with randomly coupled resonators, the property of spatial vibration transmissions, and the multi-source vibration identification results.

Based on the theoretical model, we design an actual metamaterial system with spiral-based resonators. We experimentally show the complexity of the elastic vibration propagation in the metamaterial, and verify the high uncorrelation of the vibration transmissions. Therefore, the metamaterial can be considered as the physical implementation of the measurement matrix in the compressive sensing theory. We demonstrate that the metamaterial system can accurately identify multiple broadband vibration sources with only a single sensor by combining with the compressive sensing framework. What's amazing is that the proposed metamaterial system also shows a time-dependent spatial coding ability. This means that we can track the trajectories of impact excitation events. The finding creates a new type of human-machine interaction for instruction, communication, and encryption without complex hardware and high power consumption, which shows broad application prospects in fields such as touch sensing, robot tactile sensing, collision tracking and Internet of Things.  

Figure 3: Tracking of the trajectories of vibration excitation events “SJTU” using the single-sensor identification of elastic vibration sources.

This work opens up an avenue to design and control vibration transmissions via metamaterials. The proposed metamaterial system presents a new type of compact vibration traceability device, and can be integrated with many smart devices, platforms and structures (e.g., wearable devices, quadrotor drones, robots etc.), which can be used for human-machine interaction, touch sensing, information communications, and other broad fields. In addition, the proposed metamaterial design strategy is also expected to provide a basis for designing diverse simple sensing devices for vibration and other physical information.

For more information, please see our recent publication in Nature Communications: “Randomized resonant metamaterials for single-sensor identification of elastic vibrations”.