Nowadays, urban noise emerges as a distinct threat to people’s physiological and psychological health. Previous works mainly focus on the measurement and mapping of the noise by using Wireless Acoustic Sensor Networks (WASNs) and further propose some methods that can effectively reduce the noise pollution in urban environments. In addition, the research on the combination of environmental noise measurement and acoustic events recognition are rapidly progressing. In a real-life application, there still exists the challenges on the hardware design with enough computational capacity, the reduction of data amount with a reasonable method, the acoustic recognition with CNNs, and the deployment for the long-term outdoor monitoring. In this paper, we develop a novel system that utilizes the WASNs to monitor the urban noise and recognize acoustic events with a high performance. Specifically, the proposed system mainly includes the following three stages: (1) We used multiple sensor nodes that are equipped with various hardware devices and performed with assorted signal processing methods to capture noise levels and audio data; (2) the Convolutional Neural Networks (CNNs) take such captured data as inputs and classify them into different labels such as car horn, shout, crash, explosion; (3) we design a monitoring platform to visualize noise maps, acoustic event information, and noise statistics. Most importantly, we consider how to design effective sensor nodes in terms of cost, data transmission, and outdoor deployment. Experimental results demonstrate that the proposed system can measure the urban noise and recognize acoustic events with a high performance in real-life scenarios.

Unsupervised anomalous sound detection by machines holds significant importance within the realm of industrial automation. Currently, the task of machine-based anomalous sound detection in complex industrial settings is faced with issues such as the challenge of extracting acoustic feature information and an insufficient feature extraction capability within the detection network. To address these challenges, this study proposes a machine anomalous sound detection method using the lMS spectrogram and ES-MobileNetV3 network. Firstly, the log-Mel spectrogram feature and the SincNet spectrogram feature are extracted from the raw wave, and the new lMS spectrogram is formed after fusion, serving as network input features. Subsequently, based on the MobileNetV3 network, an improved detection network, ES-MobileNetV3, is proposed in this paper. This network incorporates the Efficient Channel Attention module and the SoftPool method, which collectively reduces the loss of feature information and enhances the feature extraction capability of the detection network. Finally, experiments are conducted on the dataset provided by DCASE 2020 Task 2. Our proposed method attained an averaged area under the receiver operating characteristic curve (AUC) of 96.67% and an averaged partial AUC (pAUC) of 92.38%, demonstrating superior detection performance compared to other advanced methods.

Deep learning (DL)-based side-channel analysis (SCA) has posed a severe challenge to the security and privacy of embedded devices. During its execution, an attacker exploits physical SCA leakages collected from profiling devices to create a DL model for recovering secret information from victim devices. Despite this success, recent works have demonstrated that certain countermeasures, such as random delay interrupts or clock jitters, would make these attacks more complex and less practical in real-world scenarios. To address this challenge, we present a novel denoising scheme that exploits the U-Net model to pre-process SCA traces for “noises” (i.e., countermeasures) removal. Specifically, we first pre-train the U-Net model on the paired noisy-clean profiling traces to obtain suitable parameters. This model is then fine-tuned on the noisy-only traces collected from the attacking device. The well-trained model will be finally deployed on the attacking device to remove the noises (i.e., countermeasures) from the measured power traces. In particular, a new inductive transfer learning method is also utilized in our scheme to transfer knowledge learned from the source domain (i.e., profiling device) to the target domain (i.e., attacking device) to improve the model’s generalization ability. During our experimental evaluations, we conduct a detailed analysis of various countermeasures separately or combined and show that the proposed denoising model outperforms current state-of-the-art work by a large margin, e.g., a reduction of at least 30% in computation costs and 5× in guessing entropy.

The Multiple Signal Classification (MUSIC) algorithm has become one of the most popular algorithms for estimating the direction-of-arrival (DOA) of multiple sources due to its simplicity and ease of implementation. Spherical microphone arrays can capture more sound field information than planar arrays. The collected multichannel speech signals can be transformed from the space domain to the spherical harmonic domain (SHD) for processing through spherical modal decomposition. The spherical harmonic domain MUSIC (SHD-MUSIC) algorithm reduces the dimensionality of the covariance matrix and achieves better DOA estimation performance than the conventional MUSIC algorithm. However, the SHD-MUSIC algorithm is prone to failure in low signal-to-noise ratio (SNR), high reverberation time (RT), and other multi-source environments. To address these challenges, we propose a novel joint spherical harmonic domain beam-space MUSIC (SHD-BMUSIC) algorithm in this paper. The advantage of decoupling the signal frequency and angle information in the SHD is exploited to improve the anti-reverberation property of the DOA estimation. In the SHD, the broadband beamforming matrix converts the SHD sound pressure to the beam domain output. Beamforming enhances the incoming signal in the desired direction and reduces the SNR threshold as well as the dimension of the signal covariance matrix. In addition, the 3D beam of the spherical array has rotational symmetry and its beam steering is decoupled from the beam shape. Therefore, the broadband beamforming constructed in this paper allows for the arbitrary adjustment of beam steering without the need to redesign the beam shape. Both simulation experiments and practical tests are conducted to verify that the proposed SHD-BMUSIC algorithm has a more robust adjacent source discrimination capability than the SHD-MUSIC algorithm.

This paper considers the task of indoor subarea localization by acoustic sensing. In many practical scenarios, such as ordinary office rooms, subarea localization based on a single acoustic property without subsidiaries is always tricky due to similar space physical structures. To combat the indoor subarea localization problem without any dedicated devices, we propose a framework allowing the ordinary smartphone to quickly and easily determine its indoor subarea location. The proposed structure is to first construct the multidimensional fingerprint of each environmental background audio in a room using the auditory scene analysis method; then extract the room-level fingerprint by the acoustic background spectrum method, and cluster the subarea-level fingerprints using the Pearson correlation coefficient technology; subsequently, together with the offline stage long short-term memory recurrent neuron network and online stage least variance algorithm, the effective indoor subarea localization was achieved; finally, the localization results are presented not only in text form but also in a privacy-protected color sound pattern which is built by chroma DCT-reduced log pitch. Our experiments with real-world audios, which include 96-h environment background audios of 14 different rooms, presents that the room-level localization accuracy was higher than 90% and the subarea-level accuracy could be up to 97.64%. These results strongly proved that the proposed framework could perform well even in small complicated rooms.

In this paper, we study the joint problem of multichannel selection and data scheduling for high-frequency (HF) communication under jamming environment. Prior anti-jamming work mainly discussed fixed transmission time and considered saturated scenarios in which the agent always had packets to transmit. But HF dynamic spectrum environment and time-varying communication demand make traditional anti-jamming methods ineffective. To cope with above challenge, dynamic transmission time and packet scheduling are considered in this manuscript. The transmitter selects the working channel and the number of packets to be transmitted according to current jamming environment and buffer state. Simultaneously, channel diversity which allows the transmitter to select multiple channels to send same packet is also considered to overcome the unreliable and unstable characteristics of HF channel. We formulate the decision making process as a Markov decision process (MDP) and propose an interference-aware reinforcement learning algorithm. The proposed algorithm combining Q learning with upper confidence bounds for trees (UCT) balances exploration and exploitation of action set. Simulation results show that higher network throughput and less packet loss are obtained by the proposed algorithm.