Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jun 16;22(44):e74229. doi: 10.1002/smll.74229

Multifunctional Piezoelectric Foams Enabling Noise Absorption and Deep‐Learning‐Driven Motion Recognition

Shuyi Shen 1, Yanjun Lu 1,2, Zishi Jiang 1, Yawen Wu 1, Kaiyu Feng 1, Dou Zhang 1, Yan Zhang 1,✉
PMCID: PMC13450366  PMID: 42301901

ABSTRACT

Industrial and traffic noise pollution has emerged as a growing concern. However, conventional porous materials are restricted by high density and limited functionality. Lightweight and multifunctional acoustic materials are desirable for environmental and wearable applications. In this study, the bacterial cellulose‐Ba0.85Ca0.15Zr0.1Ti0.9O3 (BCZT)‐melamine composite foams (BZMFs) with hierarchical pore channels were fabricated through freeze‐casting. The hierarchical structure endows BZMFs with a noise reduction coefficient (NRC) of 0.24 under a low areal density of 11.36 mg·cm−2. The introduction of BCZT nanofibers imparts piezoelectric sensing functionality, achieving a sensitivity of 0.256 V·kPa−1, and the piezoelectric output remained stable after the sample completed 1890 dynamic compression cycles. Moreover, coupling BZMFs with a convolutional neural network enables accurate recognition of human motions, with classification accuracy exceeding 93.75%. These results demonstrate the multifunctionality of BZMFs and highlight their promising potential in noise reduction and intelligent motion monitoring.

Keywords: deep learning, freeze‐casting, hierarchical structure, piezoelectric functionality, sound absorption

Lightweight bacterial cellulose‐BCZT‐melamine composite foams with hierarchical pores are prepared using the freeze‐casting method. These porous materials provide both sound absorption and stable piezoelectric sensing capabilities. By combining the sensor data with a convolutional neural network, accurate human motion recognition is achieved, demonstrating practical multifunctionality in acoustics and electronics.

graphic file with name SMLL-22-e74229-g003.jpg

1. Introduction

Worldwide urbanization and industrialization have intensified the demand for transportation and industrial equipment, and the increasing use of vehicles and machinery contributes to inevitable noise pollution [1]. Noise pollution poses a threat to people's mental and physical health, including effects on insomnia [2], irritability [3], and cardiovascular diseases [4]. Meanwhile, in the era of the Internet of Things (IoT), intelligent electronic devices have been widely applied across various fields [5]. These devices can function as sensors, thereby enabling real‐time monitoring to optimize human‐environment interactions, while offering functionalities such as exercise and health monitoring [6], as well as sleep‐quality assessment, thereby contributing to improved public health [7, 8, 9]. Therefore, a combination of sound absorption and intelligence will improve the degree of functional integration, optimize space utilization, and improve portability.

Currently, cameras remain the most widely deployed sensing devices in the IoT era due to their ability to acquire images and perform real‐time monitoring. However, their reliance on visual data raises significant concerns related to privacy infringement and data security [10, 11, 12]. To overcome these limitations, capacitive sensors have emerged as a promising non‐visual alternative, capable of detecting human presence and motion without capturing images [13]. These capacitive sensors operate by detecting changes in capacitance arising from variations in the distance or dielectric constant between the electrodes. However, both cameras and capacitive sensors need an uninterrupted power supply, which increases maintenance costs and limits their long‐term applicability in large‐scale or resource‐constrained environments. As a type of self‐powered sensor, piezoelectric materials can convert mechanical energy, such as vibration, pressure, and shock, into electrical signals [14]. Among piezoelectric materials, lead zirconate titanate (PZT) piezoceramics have many applications in smart sensing [15]. Li et al. [16] developed a wearable piezoelectric sensor device based on self‐polarized PZT thin films directly grown on flexible mica substrates. The device achieves high‐fidelity arterial pulse waveform monitoring with a correlation coefficient of 0.99 relative to a commercial sensor. It further enables calibration‐free blood pressure conversion through a deep learning model. However, PZT piezoelectric ceramic sensors also have drawbacks. During the production and recycling of PZT ceramics, the lead volatilized during sintering may be inhaled by workers, causing serious health problems such as cognitive decline, anemia, kidney failure, and movement disorders [17, 18, 19]. As a result, Ba0.85Ca0.15Zr0.1Ti0.9O3 (BCZT) ceramics have attracted extensive attention for their lead‐free and excellent electrical properties [20, 21]. However, ceramic sheets generally cannot achieve effective noise reduction because of their structural characteristics [22, 23, 24].

Traditional methods of sound absorption primarily involve resonators and porous materials [25]. For resonators, sound absorption is achieved by forcing noise vibrations to match the resonance frequency, thereby converting acoustic energy into structural mechanical energy and heat [26]. Resonators are typically effective within a narrow absorption range, but are unsuitable for broadband noise [27]. Thus, their applications are limited. In contrast, porous materials used in residential and industrial buildings exhibit a broader absorption range than resonant structures [28]. Porous materials are generally classified into mineral fibers and polymer foams. However, mineral fibers are limited in application due to their low mechanical strength [25]. In addition, fragments may be detrimental to human health, causing lung fibrosis, respiratory tract inflammation, and cancer [29, 30]. Although polymer foams avoid these problems, their sound absorption performance is insufficient at low frequencies (100–2000 Hz) [26, 31]. The thickness and density of polymer materials need to be increased to meet the requirements of applications in sound‐absorbing environments, which leads to increased fuel consumption and reduced interior space [32]. The incorporation of nanophases can enhance the sound absorption performance of certain polymer‐based materials by altering their pore structures. Nine et al. [33] embed graphene oxide (GO) sheets into the melamine‐formaldehyde resin (MF) skeleton through self‐assembly and low‐temperature curing to prepare lamella structures with different densities, denoted as MFGO. The MFGO with a thickness of 6.5 ± 0.5 mm exhibited an improved sound absorption coefficient of approximately 0.17 at a frequency of 2000 Hz. However, the random porous architecture predominantly reflects incident sound waves, restricting their transmission and dissipation and thereby limiting the effective acoustic absorption performance [34].

In recent years, freeze‐casting has attracted increasing attention owing to its ability to adjust pore structures [35]. Freeze‐casting enables the construction of hierarchical pore structures, thereby forming complex 3D pore networks. The slurry is then frozen under a directional temperature gradient, causing the solvent to solidify unidirectionally along the gradient. After solidification, the sample is freeze‐dried to sublimate and remove the solvent, leaving a lamellar pore structure templated by the solvent. This structure offers advantages by creating tortuous propagation paths and attenuation possibilities. Oh et al. [36] have presented a GO‐polyurethane hybrid aerogel with parallel‐aligned and perpendicular‐aligned structures by freeze‐casting. In the low‐ and mid‐frequency ranges, the aerogels of different alignments effectively attenuated acoustic energy, achieving sound absorption coefficients of 0.57 for the parallel‐aligned structure and 0.52 for the perpendicular‐aligned structure at 2000 Hz. However, despite the success of freeze‐casting in constructing hierarchical porous structures, the reported materials cannot meet the requirements for intelligent materials. Recently, increasing efforts have been dedicated to the development of self‐powered flexible materials for multifunctional sensing applications. Nevertheless, achieving a combination of broadband sound absorption and robust piezoelectric sensing within a lightweight material remains difficult [37, 38, 39, 40]. The core challenge is the structural collapse of highly porous acoustic materials under cyclic stress, which destroys the stability of piezoelectric signals.

Herein, a composite material with a hierarchical porous structure was successfully developed by using the freeze‐casting technique, which is referred to as BZMFs (bacterial cellulose‐BCZT‐melamine composite foam). The designed hierarchical structure possesses a porosity of 99.8% and tortuosity of 3.41. This complex internal structure ensures a noise reduction coefficient (NRC) of 0.24 under a low areal density of 11.36 mg·cm−2. At the same time, the melamine skeleton supports the freeze‐casting network, which allows BZMFs to maintain structural integrity and steady piezoelectric output after 1890 dynamic compression cycles. With the help of a deep‐learning algorithm, BZMFs can be applied to the sports monitoring system. Overall, these findings indicate that BZMFs can potentially integrate acoustic and sensing functions within a single material system to resolve the stability issues of lightweight porous sensors.

2. Results and Discussion

2.1. Fabrication and Characterization of BZMFs

The preparation of BZMFs involves four main steps: preparing a suspension that thoroughly mixes BC, Ag nanowires (NWs), and BCZT nanofibers (NFs); vacuum impregnating the melamine foam with this suspension; directionally freezing the foam; and then freeze‐drying the obtained frozen samples (Figure 1a). SEM images of the Ag NWs (Figure S2), BCZT NFs (Figure S3), and BCZT NFs within the sample (Figure S4) are presented. Besides, FTIR spectra of B0.2Z5MF and melamine foam samples are shown in Figure S5. As shown in Figure 1b, the BZMFs samples were lightweight; a sample with a volume of 152π × 10 mm3 could be supported by petals. In addition, Figure S6 shows a photograph of the sample fabricated into the shape of the letters “CSU”. Samples with different BC and BCZT solid concentrations (0.05/5 wt.%, 0.1/5 wt.%, 0.2/0, 0.2/2.5 wt.%, 0.2/5 wt.%, 0.2/7.5 wt.%) were prepared and marked as B0.05Z5MF, B0.1Z5MF, B0.2Z0MF, B0.2Z2.5MF, B0.2Z5MF and B0.2Z7.5MF respectively. In addition, samples (rBZMFs) with random channels were prepared in a homogeneous cold source without a temperature gradient.

FIGURE 1.

FIGURE 1

Open in a new tab

Preparation, characterization, and application examples of BZMFs. (a) Schematic diagram of the design and preparation process of BZMFs. (b) Image of 152π × 10 mm3 BZMFs supported by petals. (c) SEM images of B0.2Z5MFs, showing a hierarchical structure of oriented and random channels. (d) Schematic diagram of BZMFs as sound‐absorbing devices and self‐powered sensors used in the assembly of smart earmuffs; Schematic of a convolutional neural network for machine learning; Data graphs from BZMFs monitoring five different movements in smart earmuffs; Schematic diagram of BZMFs monitoring movement in smart earmuffs.

Scanning electron microscopy (SEM) images of B0.2Z5MF frozen under a controlled temperature gradient are presented in Figure 1c. During directional freezing, the suspension gradually transformed into ice crystals, which propagated along the melamine fiber scaffold. Subsequently, BC was observed to form oriented channels along the freezing direction, exhibiting parallel and aligned features. These oriented channels were found to interconnect with generated channels within the melamine matrix and adjacent cellulose fibers, forming flocculent fiber structures. The SEM images of BZMF samples with 0.1% and 0.05% BC solid contents are shown in Figure S7a,b, respectively. Additionally, the honeycomb‐like random channels of the melamine matrix integrate with the ice‐templated aligned BC channels, forming an interwoven hierarchical porous network. To further quantify the internal pore architecture, the composite foams with different solid contents were characterized by mercury intrusion porosimetry (MIP), as shown in Figure S8. For BC content 0.2%, the porosity value was 99.8%, indicating a highly open and lightweight framework. As the solid content increased, the pore channels became more complex and tortuous, leading to an increase in tortuosity up to 3.41.

Meanwhile, BZMFs were found to be capable of generating piezoelectric signals for the smart earmuffs. Figure 1d illustrates that BZMFs integrated with a deep‐learning algorithm show potential for applications in motion monitoring.

2.2. Sound Absorption Properties of BZMFs

Figure 2a–c compares the microstructure and sound absorption performance of BZMFs (B0.2Z5MF) prepared with the same solid content but different freezing methods. Figure 2a shows that BC grows into cotton‐like fibers that adhere to the melamine foam skeleton under directional freezing, whereas under uniform cooling, it disperses as random channels surrounding the skeleton. As shown in Figure 2b, the directional channels exhibit efficient sound absorption. At a frequency of 2500 Hz, the sound absorption coefficient of the 10‐mm sample reached 0.63, which is 16.6% higher than the sound absorption coefficient of rBZMFs with random pores. Figure S9 further indicates that the NRC of the directional channels is 4.35% higher than that of the random channels. Compared with the random channels formed by BC, the directional channels generate tortuous pathways and facilitate more efficient conversion of acoustic energy into thermal energy [41, 42, 43]. Additionally, the influence of different freezing temperature gradients on the acoustic performance was investigated in Figure 2c to demonstrate the validity of the fabrication process. The NRC and average sound absorption coefficients were also calculated and are shown in Figure S10.

FIGURE 2.

FIGURE 2

Open in a new tab

Sound absorption properties and mechanisms of BZMFs. (a) SEM images of BZMFs and rBZMFs. (b) Absorption coefficients of B0.2Z5MF with different channels. (c) Absorption coefficients of B0.2Z5MF with different freeze‐casting temperatures. Sound absorption mechanisms of (d) BZMFs with oriented channels and (e) BZMFs with random channels. (f) Absorption coefficients of samples with different BC solid contents. (g) Noise reduction coefficients and average absorption coefficients of BZMFs with different BC solid contents samples. (h) Comparison of noise reduction performance of B0.2Z5MF and other acoustic materials [47, 48, 49, 50].

Figure 2d,e illustrates the acoustic schematics of BZMFs in different channel directions. As shown in Figure 2d, in BZMFs with oriented pore channels, incident sound waves are converted into thermal energy during propagation between aligned pores, while some of the waves are reflected at the surface or within the framework, and a residual fraction transmits through the material [25, 44, 45]. By contrast, Figure 2e demonstrates that in rBZMFs with randomly oriented channels, the irregular pathways cause scattering and reflection, which shorten the effective propagation distance and limit energy dissipation [39, 46]. Consequently, fewer sound waves are absorbed, and more are reflected or transmitted [40]. This mechanism explains the sound absorption of oriented channels over random channels, as observed in Figure 2b,c. Because of this advantage, directional freezing was employed in subsequent experiments to prepare BZMFs with optimized acoustic properties.

The influence of BC solid content on the acoustic properties of BZMFs is shown in Figure 2f–h. Increasing the BC solid content from 0.05 to 0.2 wt.% gradually enhances the sound absorption coefficients at frequencies between 1000 and 2500 Hz, with the sound absorption coefficient of the 10‐mm‐thick B0.2Z5MF at 2500 Hz increased by 73.2% and 18.5% compared with that of B0.05Z5MF and B0.1Z5MF (Figure 2f). This is attributed to the formation of more tortuous propagation channels, which extend the travel path of sound waves and increase energy dissipation [34]. As shown in Figure 2g, B0.2Z5MF exhibits the efficient acoustic performance, with NRC and average absorption coefficients of 0.24 and 0.61, respectively, corresponding to enhancements of 96.7% and 119% compared with melamine foam. Figure S11 illustrates the effect of thickness on the acoustic performance of the B0.2Z5MF, including its sound absorption coefficients, NRC, and average sound absorption. In addition, the effect of different BCZT contents on the sound absorption performance of the composite foam was investigated and presented in Figure S12. Furthermore, Figure 2h demonstrates that BZMFs maintain the highest NRC (0.24) at a relatively low areal density (11.36 mg·cm−2), underscoring their lightweight but enhanced noise reduction performance. This is primarily ascribed to the composite internal channel structure generated by increased BC solid contents, which intensifies air friction and viscous flow against the wall surfaces during propagation, thereby facilitating efficient sound energy dissipation.

To evaluate the practical noise reduction capability of BZMFs, their performance was compared with that of commercial melamine foam under noise environments that are characterized by either predominantly low‐frequency or predominantly high‐frequency components. Figure 3a,b presents the frequency distributions of noise generated by a hair dryer and an alarm clock, respectively. As shown in Figure 3a, the hair dryer spectrum is concentrated below 7000 Hz, with dominant components below 3000 Hz, indicating that it primarily produces low‐to‐mid‐frequency noise. In contrast, Figure 3b demonstrates that the alarm clock mainly produces high‐frequency noise, and the primary noise frequency band is concentrated at 3000 Hz and above. As illustrated in Figure 3c, B0.2Z5MF reduced the hair dryer noise from 106.6 dB to 83.3 dB, whereas melamine foam achieved only a 12.2 dB reduction. Moreover, for high‐frequency noise from the alarm clock, B0.2Z5MF achieved an attenuation of 19.9 dB, corresponding to 1.76 times the reduction of melamine foam.

FIGURE 3.

FIGURE 3

Open in a new tab

Practical noise reduction tests of BZMF. Spectrum of (a) low‐frequency noise (hair dryer) and (b) high‐frequency noise (alarm clock). Noise reduction test and performance of B0.2Z5MF for (c) low‐frequency noise situation and (d) high‐frequency noise situation.

2.3. Piezoelectric and Mechanical Properties of BZMFs

Here, the electrical performance of BZMF samples was investigated. The mechanism of electrical signal generation in the sample is illustrated in Figure S13. The electrical output of BZMFs with incorporated Ag NWs was compared to that of pristine samples, which are B0.2Z0.5MF without Ag NWs, under the same strike frequency and pressure. As shown in Figure 4a, the incorporation of Ag NWs significantly enhances the output, yielding ∼2.1 compared to −0.74 V for the original sample (B0.2Z0.5MF without Ag NWs). This improvement is mainly attributed to the conductive Ag NWs network, which not only facilitates charge transport and reduces interfacial resistance [51, 52, 53] but also concurrently promotes local polarization to enhance the piezoelectric response of the foam [54, 55]. Figure 4b shows the voltage signal tests on B0.2Z5MF under various pressures (0.8, 4.0, 8.0, 12.0 and 14.4 kPa). The results revealed that the output voltage increased gradually, reaching 3.5 V at 14.4 kPa. Furthermore, Figure 4d demonstrates a linear correlation between pressure and electrical output, with a coefficient of determination (R2 = 0.985) and sensitivity (the slope of the straight line in Figure 4d [56, 57]) of 0.256 V·kPa−1. Figure 4c illustrates the variation in peak voltage output at different strike frequencies, where the electrical output rose with increasing frequency, reaching ∼2.85 V at 2.0 Hz. Meanwhile, to evaluate the piezoelectric cycling stability of the sample, B0.2Z5MF was subjected to 1890 loading cycles under a pressure of 4.8 kPa at a frequency of 2 Hz, as shown in Figure 4e. Throughout the cyclic testing process, the voltage output remained stable, with both the initial 10 cycles and the final 10 cycles averaging ∼1.22 V, demonstrating the electromechanical stability and durability of the piezoelectric foam. These results illustrate that the embedded BCZT NFs in BZMFs function as self‐powered piezoelectric elements, enabling the composites to generate stable electrical signals under external stress.

FIGURE 4.

FIGURE 4

Open in a new tab

Voltage output signals and mechanical properties of BZMFs. (a) Comparison of piezoelectric signals of B0.2Z5MF with and without Ag NWs. (b) Voltage output of B0.2Z5MF subjected to different pressures. (c) Voltage output of B0.2Z5MF struck at different frequencies. (d) Linear fitting of pressure to voltage signal. (e) The 1890 times piezoelectric cycles of sample B0.2Z5MF under a pressure of 4.8 kPa and a frequency of 2 Hz. (f) Compressive stress‐strain curves of BZMFs. (g) Compressive stress‐strain curves of B0.2Z5MF over 155 cycles at 60% strain. (h) Optical images of the compression‐recovery process. (i) Changes of energy dissipation coefficients and maximum stress during compression cycles. (j) Changes of load on B0.2Z5MF during compression cycles.

In addition, the mechanical properties are crucial for practical applications. As illustrated in Figure 4f, the samples exhibited a closed‐loop compressive stress‐strain curve. The stress correspondingly rose as the strain increased, reaching a maximum of 40 kPa at 80% strain, after which the sample exhibited recovery behavior. The interconnected network of BC and the melamine foam skeleton endowed BZMFs with excellent compressive strength. To further evaluate the durability of BZMFs, the cyclic compression tests were conducted, as shown in Figure 4g,j. Figure 4g presents the compressive stress‐strain curves over loading‐unloading cycles, where the stress response gradually stabilized after repeated compression, indicating structural resilience and fatigue resistance. The rebound process shown in Figure 4h further confirms the recovery capability of BZMFs during dynamic deformation. Figure 4j further shows the corresponding load values during cyclic compression, exhibiting stable responses with average loads of 10.88, 9.81 and 9.53 N at different stages of the cycling process. The average load on the specimen during the final 10 cycles remained at 87.59% of the initial average load observed during the first 10 cycles, confirming its long‐term mechanical stability. The changes in max stress, energy dissipation coefficients, and Young's modulus of B0.2Z5MF specimens with increasing cycle numbers are shown in Figure S14. As shown in Figure 4i, the energy dissipation coefficients and maximum compressive stress both rose with increasing BC content, as the interconnected BC network became reinforced. Besides, the shape retention of samples without a melamine foam skeleton collapsed completely under 50% compression (Figure S15), showing no practical usability [58, 59]. These results demonstrate the efficient sound‐absorbing, mechanical and stable piezoelectric properties of BZMFs, making them promising candidates for multifunctional materials when compared with representative literature (Table S2).

2.4. Sensing Applications of BZMFs

To assess the sensing potential of BZMFs, a smart earmuff system was designed. The B0.2Z5MF sample was affixed to the earmuff shell, and the assembled device was subsequently worn by test subjects to record piezoelectric signals during various motion tasks, as shown in Figure S16. Five different movements (walking, shooting, stretching, running and sitting & standing) were designed to validate the application of BZMFs in actual motion monitoring. As shown in Figure 5a–e, the electrical signals generated by each movement are independent compared with the other four actions, with each movement characterized by stable and repeatable waveforms. Figure 5f presents a representative test, in which a 2.5 × 2.5 cm2 B0.2Z5MF sample was affixed to an earmuff for evaluation.

FIGURE 5.

FIGURE 5

Open in a new tab

Sensing signals of BZMFs under five different actions and application in action recognition based on the CNN model. Sensing signals and schematics for (a) walking action, (b) shooting action, (c) stretching action, (d) running action and (e) sitting & standing actions. (f) optical diagram of testing system (I. test earmuffs and samples of the installation, II—VI. five types of action of the actual demonstration). (g) The construction process of the CNN model. (h) The accuracy and loss function of the CNN model after 900 epochs. (i) The F1‐scores for the five actions. (j) The confusion matrix for the five actions.

To further validate the feasibility of BZMF‐based sensors for motion monitoring, a 1D convolutional neural network (CNN) was constructed for the classification of five representative human activities: walking, shooting, stretching, running and sitting & standing (Figure 5g). This AI‐driven data processing is efficient for smart sensing systems [60]. The network architecture consisted of two convolutional layers to extract features from the sensing signals, followed by rectified linear unit (ReLU) activations to enhance feature nonlinearity [61, 62]. Max‐pooling layers were adopted to reduce the dimensionality while retaining key feature representations [63]. The fully connected layers integrated these extracted features, and a softmax function was applied in the output layer to assign the signals into the five motion categories [64].

As shown in Figure 5h, the validation accuracy exceeded 93.75% after 900 epochs, while the validation loss decreased consistently. Figure 5i illustrates that the F1‐scores across all categories exceeded 0.9, indicating that the model achieved balanced precision and recall for different motion types. Moreover, the confusion matrix (Figure 5j) demonstrates classification performance.

These results highlight the synergy between BZMF‐based flexible sensors and deep learning algorithms for accurate motion recognition. CNNs provide enhanced feature extraction, enabling reliable identification of five signal variations. Such integration holds promise for practical applications in intelligent sports monitoring, rehabilitation assessment, fatigue warning, and fall detection in elderly care.

3. Conclusion

BZMFs with hierarchical pore channels were fabricated via freeze‐casting. This unique architecture enables efficient sound absorption while maintaining a low areal density of 11.36 mg·cm−2. The hierarchical channels enhance broadband acoustic absorption (NRC = 0.24), whereas the incorporation of BCZT NFs imparts piezoelectric sensing capabilities with a sensitivity of 0.256 V·kPa−1. Additionally, the piezoelectric signal remained stable after the sample cycled through 1890 dynamic compression cycles. Furthermore, integration with a convolutional neural network (CNN) enables accurate action recognition with an accuracy exceeding 93.75%. With the inclusion of more diverse motion datasets, BZMFs combined with CNNs hold great potential for intelligent applications such as precise training optimization, rehabilitation assessment, and fall detection.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll74229‐sup‐0001‐SuppMat.docx.

SMLL-22-e74229-s001.docx (5.1MB, docx)

Acknowledgements

The authors acknowledge the support of the Advanced Materials‐National Science and Technology Major Project (2025ZD0619902), Xiaomi Young Talents Program and the State Key Laboratory of Powder Metallurgy, Central South University, Changsha, China.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  • 1. Welch D., Shepherd D., Dirks K. N., and Reddy R., “Health Effects of Transport Noise,” Transport Reviews 43, no. 6 (2023): 1190–1210. [Google Scholar]
  • 2. Basner M. and McGuire S., “WHO Environmental Noise Guidelines for the European Region: a Systematic Review on Environmental Noise and Effects on Sleep,” International Journal of Environmental Research and Public Health 15, no. 3 (2018): 519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Hahad O., Kuntic M., Al‐Kindi S., et al., “Noise and Mental Health: Evidence, Mechanisms, and Consequences,” Journal of Exposure Science & Environmental Epidemiology 35, no. 1 (2025): 16–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Münzel T., Sorensen M., and Daiber A., “Transportation Noise Pollution and Cardiovascular Disease,” Nature Reviews Cardiology 18, no. 9 (2021): 619–636. [DOI] [PubMed] [Google Scholar]
  • 5. Elgazzar K., Khalil H., Alghamdi T., et al., “Revisiting the Internet of Things: New Trends, Opportunities and Grand Challenges,” Frontiers in the Internet of Things 1 (2022): 1073780. [Google Scholar]
  • 6. Ni Y. M., Zang X. R., Yang Y., et al., “Environmental Stability Stretchable Organic Hydrogel Humidity Sensor for Respiratory Monitoring with Ultrahigh Sensitivity,” Advanced Functional Materials 34, no. 38 (2024): 12. [Google Scholar]
  • 7. Zhang H., Zhang R., and Sun J., “Developing Real‐Time IoT‐Based Public Safety Alert and Emergency Response Systems,” Scientific Reports 15, no. 1 (2025): 29056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Fraga‐Lamas P., Fernandez‐Carames T. M., Suarez‐Albela M., Castedo L., and Gonzalez‐Lopez M., “A Review on Internet of Things for Defense and Public Safety,” Sensors 16, no. 10 (2016): 1644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Vaghasiya J. V., Mayorga‐Martinez C. C., and Pumera M., “Wearable Sensors for Telehealth Based on Emerging Materials and Nanoarchitectonics,” Npj Flexible Electronics 7, no. 1 (2023): 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Ahmed A. A. and Nyarko B., “Smart‐Watcher: an AI‐Powered IoT Monitoring System for Small‐Medium Scale Premises,” 2024 International Conference on Computing, Networking and Communications (ICNC) (2024), 139–143.
  • 11. Shan C., Brea V. M., and Velipasalar S., “Special Issue on Smart Cameras for Real‐Time Image and Video Processing,” Journal of Real‐Time Image Processing 17, no. 6 (2020): 1755–1756. [Google Scholar]
  • 12. Shen Y. Z., Shen S. G., Li Q., Zhou H. P., Wu Z. D., and Qu Y. Y., “Evolutionary Privacy‐Preserving Learning Strategies for Edge‐Based IoT Data Sharing Schemes,” Digital Communications and Networks 9, no. 4 (2023): 906–919. [Google Scholar]
  • 13. Valtonen M., Vuorela T., Kaila L., and Vanhala J., “Capacitive Indoor Positioning and Contact Sensing for Activity Recognition in Smart Homes,” Journal of Ambient Intelligence and Smart Environments 4, no. 4 (2012): 305–334. [Google Scholar]
  • 14. Cao X. L., Xiong Y., Sun J., Zhu X. X., Sun Q. J., and Wang Z. L., “Piezoelectric Nanogenerators Derived Self‐Powered Sensors for Multifunctional Applications and Artificial Intelligence,” Advanced Functional Materials 31, no. 33 (2021): 2102983. [Google Scholar]
  • 15. Dagdeviren C., Su Y. W., Joe P., et al., “Conformable Amplified Lead Zirconate Titanate Sensors with Enhanced Piezoelectric Response for Cutaneous Pressure Monitoring,” Nature Communications 5 (2014): 10. [DOI] [PubMed] [Google Scholar]
  • 16. Li M., Aoyama J., Inayoshi K., and Zhang H., “Wearable PZT Piezoelectric Sensor Device for Accurate Arterial Pressure Pulse Waveform Measurement,” Advanced Electronic Materials 11, no. 9 (2025): 2400852. [Google Scholar]
  • 17. Boskabady M., Marefati N., Farkhondeh T., Shakeri F., Farshbaf A., and Boskabady M. H., “The Effect of Environmental Lead Exposure on human Health and the Contribution of Inflammatory Mechanisms, a Review,” Environment International 120 (2018): 404–420. [DOI] [PubMed] [Google Scholar]
  • 18. Kumar A., Amit K., Cabral‐Pinto M. M. S., et al., “Lead Toxicity: Health Hazards, Influence on Food Chain, and Sustainable Remediation Approaches,” International Journal of Environmental Research and Public Health 17, no. 7 (2020): 2179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Lee M. H., Kim D. J., Park J. S., et al., “High‐Performance Lead‐Free Piezoceramics with High Curie Temperatures,” Advanced Materials 27, no. 43 (2015): 6976–6982. [DOI] [PubMed] [Google Scholar]
  • 20. Amorin H., Venet M., Garcia J. E., et al., “Insights into the Early Size Effects of Lead‐Free Piezoelectric Ba0.85Ca0.15Zr0.1Ti0.9O3 ,” Advanced Electronic Materials 10, no. 2 (2024): 2300556. [Google Scholar]
  • 21. Shi C., Yan F., Ge G., Wei Y., Zhai J., and Yao W., “Significantly Enhanced Energy Storage Performances and Power Density in (1 − x)BCZT‐xSBT Lead‐Free Ceramics via Synergistic Optimization Strategy,” Chemical Engineering Journal 426 (2021): 130800. [Google Scholar]
  • 22. Wang W., Liu H., and Gu W., “A Novel Fabrication Approach for Improving the Mechanical and Sound Absorbing Properties of Porous Sound‐Absorbing Ceramics,” Journal of Alloys and Compounds 695 (2017): 2477–2482. [Google Scholar]
  • 23. Ji X., Wang C. B., Li S. B., et al., “Structural and Electrical Properties of BCZT Ceramics Synthesized by Sol–gel Process,” Journal of Materials Science: Materials in Electronics 29, no. 9 (2018): 7592–7599. [Google Scholar]
  • 24. Lou J. Y., He C., Shui A. Z., and Yu H. L., “Enhanced Sound Absorption Performance of Porous Ceramics with Closed‐Pore Structure,” Ceramics International 49, no. 23 (2023): 38103–38114. [Google Scholar]
  • 25. Cao L., Fu Q., Si Y., Ding B., and Yu J., “Porous Materials for Sound Absorption,” Composites Communications 10 (2018): 25–35. [Google Scholar]
  • 26. Li P. P., Zhang X., Li Y., Zhao C. Y., Yu J. Y., and Si Y., “High‐Performance Gate‐All‐Around Field Effect Transistors Based on Orderly Arrays of Catalytic Si Nanowire Channels,” Nano‐Micro Letters 18, no. 1 (2025): 46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Xu J. S., Ye Y. F., Dong T., et al., Advancement of Science 12, no. 27 (2025): 25. [Google Scholar]
  • 28. Vasina M., Mesicek J., Manek M., Ma Q.‐P., Hajnys J., and Petru J., “Investigation of Factors Affecting the Sound Absorption Behaviour of 3D Printed Hexagonal Prism Lattice Polyamide Structures,” Scientific Reports 14, no. 1 (2024): 30852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Lippmann M., “Effects of fiber Characteristics on Lung Deposition, Retention, and Disease,” Environmental Health Perspectives 88 (1990): 311–317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Gualtieri A. F., “Journey to the Centre of the Lung. The Perspective of a Mineralogist on the Carcinogenic Effects of Mineral Fibres in the Lungs,” Journal of Hazardous Materials 442 (2023): 130077. [DOI] [PubMed] [Google Scholar]
  • 31. Du Z. P., Yao D. X., Xia Y. F., et al., “Highly Porous Silica Foams Prepared via Direct Foaming with Mixed Surfactants and Their Sound Absorption Characteristics,” Ceramics International 46, no. 9 (2020): 12942–12947. [Google Scholar]
  • 32. Zong D. D., Cao L. T., Yin X., et al., “Pan‐Cancer Circulating Tumor DNA Detection in over 10,000 Chinese Patients,” Nature Communications 12, no. 1 (2021): 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Nine M. J., Ayub M., Zander A. C., Tran D. N. H., Cazzolato B. S., and Losic D., “Graphene Oxide‐Based Lamella Network for Enhanced Sound Absorption,” Advanced Functional Materials 27, no. 46 (2017): 1703820. [Google Scholar]
  • 34. Xiong J. K., Liu J. L., Lin W. G., et al., “Enhanced Broadband Acoustic Absorption in Commercial Foam via Multiwall Carbon Nanotube‐Induced Pore Reconstruction,” Advancement of Science 12, no. 19 (2025): 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Lee H., Jeon J.‐W., Koh Y.‐H., and Kim H.‐E., “Dual‐Scale Porosity Alumina Structures Using Ceramic/Camphene Suspensions Containing Polymer Microspheres,” Materials 15, no. 11 (2022): 3875. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Oh J. H., Kim J., Lee H., Kang Y., and Oh I. K., “Directionally Antagonistic Graphene Oxide‐Polyurethane Hybrid Aerogel as a Sound Absorber,” ACS Applied Materials & Interfaces 10, no. 26 (2018): 22650–22660. [DOI] [PubMed] [Google Scholar]
  • 37. Jiang X. D., Guo J. B., Zhang Y. H., et al., “Electrophoretic Deposition of Ag‐Cu‐CTS Coatings on Porous Titanium with Photothermal‐Responsive Antibacterial Effect,” Journal of Colloid and Interface Science 682 (2025): 1116–1126. [DOI] [PubMed] [Google Scholar]
  • 38. Deng L., Zhang Y., Xu H., et al., “Wearable Fabrics Against Ultra‐Broadband Electromagnetic Interference,” Matter 9 (2026): 102697. [Google Scholar]
  • 39. Shen S., Jiang Z., Lu Y., et al., “Elastic Aerogels Prepared by Cyclic Freeze Casting for Broadband Sound Absorption and Sleep Health Monitoring,” Nano Energy 147 (2026): 111591. [Google Scholar]
  • 40. Shen S., Zhang Y., Guo W., et al., “Hierarchically Piezoelectric Aerogels for Efficient Sound Absorption and Machine‐Learning‐Assisted Sensing,” Advanced Functional Materials 34, no. 42 (2024): 2406773. [Google Scholar]
  • 41. Cao L. T., Si Y., Yin X., Yu J. Y., and Ding B., “Ultralight and Resilient Electrospun Fiber Sponge with a Lamellar Corrugated Microstructure for Effective Low‐Frequency Sound Absorption,” ACS Applied Materials & Interfaces 11, no. 38 (2019): 35333–35342. [DOI] [PubMed] [Google Scholar]
  • 42. Nine M. J., Ayub M., Zander A. C., Tran D. N. H., Cazzolato B. S., and Losic D., “Graphene Oxide‐Based Lamella Network for Enhanced Sound Absorption” Advanced Functional Materials 27, no. 46 (2017): 1703820. [Google Scholar]
  • 43. Yang M. and Sheng P., “Sound Absorption Structures: from Porous Media to Acoustic Metamaterials,” Annual Review of Materials Research 47 (2017): 83–114. [Google Scholar]
  • 44. Dong S., Duan Y., Chen X., et al., “Recent Advances in Preparation and Structure of Polyurethane Porous Materials for Sound Absorbing Application,” Macromolecular Rapid Communications 45, no. 14 (2024): 2400108. [DOI] [PubMed] [Google Scholar]
  • 45. Chen K. H., Wang D., Du J. W., et al., “Excellent Broadband Sound Absorption in Composites Manufactured by Embedding Piezoelectric Polymer Aerogels in Porous Ceramics,” Journal of Colloid and Interface Science 680 (2025): 537–545. [DOI] [PubMed] [Google Scholar]
  • 46. Shen L., Zhang H. R., Lei Y. Z., Chen Y., Liang M., and Zou H. W., “Simultaneous Adsorption of Heavy Metals and Organic dyes by β‐Cyclodextrin‐Chitosan Based Cross‐Linked Adsorbent,” Carbohydrate Polymers 255 (2021): 117486. [DOI] [PubMed] [Google Scholar]
  • 47. Berardi U. and Iannace G., “Predicting the Sound Absorption of Natural Materials: Best‐Fit Inverse Laws for the Acoustic Impedance and the Propagation Constant,” Applied Acoustics 115 (2017): 131–138. [Google Scholar]
  • 48. Na Y. and Cho G., “Sound Absorption and Viscoelastic Property of Acoustical Automotive Nonwovens and their Plasma Treatment,” Fibers and Polymers 11, no. 5 (2010): 782–789. [Google Scholar]
  • 49. Yilmaz N. D., Banks‐Lee P., Powell N. B., and Michielsen S., “Effects of Porosity, fiber Size, and Layering Sequence on Sound Absorption Performance of Needle‐Punched Nonwovens,” Journal of Applied Polymer Science 121, no. 5 (2011): 3056–3069. [Google Scholar]
  • 50. Han F. S., Seiffert G., Zhao Y. Y., and Gibbs B., “Acoustic Absorption Behaviour of an Open‐Celled Aluminium Foam,” Journal of Physics D: Applied Physics 36, no. 3 (2003): 294–302. [Google Scholar]
  • 51. Dudem B., Kim D. H., Bharat L. K., and Yu J. S., “Highly‐Flexible Piezoelectric Nanogenerators with Silver Nanowires and Barium Titanate Embedded Composite Films for Mechanical Energy Harvesting,” Applied Energy 230 (2018): 865–874. [Google Scholar]
  • 52. Lin Y., Fang T., Bai C., et al., “Ultrastretchable Electrically Self‐Healing Conductors Based on Silver Nanowire/Liquid Metal Microcapsule Nanocomposites,” Nano Letters 23, no. 23 (2023): 11174–11183. [DOI] [PubMed] [Google Scholar]
  • 53. Wen D. L., Huang P., Li B. Y., et al., “Silk Fibroin/Ag Nanowire‐Based Multifunctional Sensor for Wearable Self‐Powered Wireless Multi‐Sensing Microsystems,” Nano Energy 113 (2023): 108569. [Google Scholar]
  • 54. Shin J., Kim K., Min I. S., et al., “A Wireless Wearable Sensor System Based on a Silver Nanowire‐Decorated Silicon Nanomembrane for Precise and Continuous Hazardous Gas Monitoring,” Advanced Functional Materials 35, no. 15 (2025): 202419110. [Google Scholar]
  • 55. Zhao J., Zheng M. P., Gao X., Zhu M. K., and Hou Y. D., “High‐Performance Flexible Piezocomposites Based on an Enhanced Interfacial Polarization Effect via BCZT@Ag Heterostructure Design,” Journal of Alloys and Compounds 970 (2024): 172702. [Google Scholar]
  • 56. Huang X. D., Ma Z. Y., Xia W. T., et al., “Thickness‐Derived Optical–Electrical management in Sn‐Based Perovskite Solar Cells,” Nano Energy 129 (2024): 8. [Google Scholar]
  • 57. Zhen L. Y., Cui M., Bai X. Y., et al., “Thin, Flexible Hybrid‐Structured Piezoelectric Sensor Array with Enhanced Resolution and Sensitivity,” Nano Energy 131 (2024): 110188. [Google Scholar]
  • 58. Chen L., Chen H., Ji C., Wang Y., and Yang L., “Bacterial Cellulose Nanofibers‐Assisted Construction of Core‐Shell Structured Polyaniline Aerogel for Superior Electromagnetic Wave Absorption,” Carbohydrate Polymers 352 (2025): 123239. [DOI] [PubMed] [Google Scholar]
  • 59. Yang C., Chen C., Dong X., et al., “Ultra‐Flexible Heterogeneous‐Structured Hollow‐Fe3O4@MXene/Silver Nanowire/Bacterial Cellulose Composite Film for Wide‐Band Electromagnetic Interference Shielding Applications,” Small 21, no. 44 (2025): 202508856. [DOI] [PubMed] [Google Scholar]
  • 60. Liang Z., Deng Y., Shi Z., et al., “AI‐Driven Design of Powder‐Based Nanomaterials for Smart Textiles: from Data Intelligence to System Integration,” Advanced Powder Materials 5, no. 1 (2026): 100356. [Google Scholar]
  • 61. Kiranyaz S., Avci O., Abdeljaber O., Ince T., Gabbouj M., and Inman D. J., “1D Convolutional Neural Networks and Applications,” Journal of Physics D Applied Physics 151 (2021): 107398. [Google Scholar]
  • 62. Krizhevsky A., Sutskever I., and Hinton G. E., “ImageNet Classification with Deep Convolutional Neural Networks,” Communications of the ACM 60, no. 6 (2017): 84–90. [Google Scholar]
  • 63. Scherer D., Mueller A., and Behnke S., “Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition,” 20th International Conference on Artificial Neural Networks , (Springer, 2010): 92–101. [Google Scholar]
  • 64. Lecun Y., Bottou L., Bengio Y., and Haffner P., “Gradient‐Based Learning Applied to Document Recognition,” Proceedings of the IEEE 86, no. 11 (1998): 2278–2324. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting File: smll74229‐sup‐0001‐SuppMat.docx.

SMLL-22-e74229-s001.docx (5.1MB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Small (Weinheim an Der Bergstrasse, Germany) are provided here courtesy of Wiley

RESOURCES