Acoustic Events Research Articles

Racing in Kart Dromes: Laboratory and Site Assessment of Noise Levels from Competition and Rental Karts

Kart racing is one of the hobbies that people get passionate about from a young age. Kart dromes are commonly built in suburban or rural areas, generally surrounded by industrial zones and sporadic residential buildings. The circuits are primarily active during summer, hosting races that often extend into the evening and night hours, where each race has a duration of 20 min. This study examines the noise generated by kart dromes through acoustic measurements conducted at a kart drome located in southern Italy, where a microphone was placed at the side of the circuit for short periods in addition to a survey conducted at the nearest sensitive receptor. Another survey was conducted within the kart drome for a long-term period to record all of the variations in noise levels of a typical summer day when the races are organized during the nighttime; for this type of data, the hourly average values were taken for one week in June, July, and August, highlighting the increasing trend in the noise levels due to the kart races. However, a detailed analysis of noise emissions during different phases of kart operation revealed two significant acoustic events, such as the acceleration of pass-by peaks centered on high frequencies and strong breaking noise at curves that are centered at low-medium frequencies, causing a whistling noise of the wheels while turning the kart. This paper highlights the increasing trend in noise levels during summer nighttime races, compares on-site measurements with laboratory data, and discusses the implications for local communities and noise regulations.

A deep learning model for detecting and classifying multiple marine mammal species from passive acoustic data

Underwater passive acoustics is used worldwide for multi-year monitoring of marine mammals. Yet, the large amount of audio recordings raises the need to automate the detection of acoustic events. For instance, the increasing number of Offshore Wind Farms (OWF) raises key environmental and societal issues relating to their impacts on wildlife. In this context, monitoring marine mammals along with information on their acoustic environment throughout the OWF life cycle is crucial. The objective of this study is to evaluate the ability of a single deep learning model to precisely detect and localize, in time and in frequency, the marine mammal sounds over a wide frequency range and classify them by species and sound types.A broadband hydrophone, deployed at the Fécamp OWF (Normandy, France), recorded the underwater soundscape including sounds from marine mammals occurring in the area. To visualize these sounds, 15-s spectrograms were computed. From these images, dolphin (D) and porpoise (P) sounds were manually annotated, including different types of sounds: Click-Trains (DCT, PCT), Buzzes (DB, PB) and Whistles (DW). The spectrograms were then split into five-fold cross-validation datasets, each containing one half of manual annotations and one half of only background noise. A Faster R-CNN model was trained to precisely detect and classify the marine mammal sounds in the spectrograms.Three model output configurations were used: (1) overall detection of marine mammals (presence vs. absence), (2) detection and classification of species (two classes: dolphin, porpoise) and (3) sound types (five classes: DCT, DB, DW, PCT, PB). For the simplest configuration (1) 15.4 % of the spectrogram dataset had detections while missing only 6.6 % of annotated spectrograms. For the more precise configurations, (2) and (3), the mean Average Precision (mAP) achieved were 92.3 % (2) and 84.3 % (3), and the macro average Area under the curve (AUC) 95.7 % (2) and 94.9 % (3).This model will help to speed up the annotation processes, by reducing the spectrogram quantity to be manually analyzed and having time-frequency boxes already drawn. Several model parameters can be adjusted to trade off missed detections and false positives which need to be carefully considered and adapted to the problem. For instance, these adjustments would be particularly relevant depending on the human resources available to manually check the model detections and the criticality of missing marine mammal sounds. These models are promising, ranging from the simple detection of marine mammal presence to precise ecological inferences over the long term.

Drilling pressure relief under confining pressure: energy effect and numerical simulation analysis

ABSTRACT Mine pressure bump can be effectively mitigated through drilling pressure relief, with the success of this method largely dependent on the selection of appropriate drilling parameters. This study investigates the pressure relief behaviour of rock masses subjected to various drilling parameters, accounting for lateral constraints and vertical stress, and employing independently designed drilling testing equipment. A combination of theoretical research, laboratory experiments and numerical simulations was employed to derive the findings. The results indicate that increases in borehole size and depth lead to a decrease in both the peak strength of the rock mass and the accumulated elastic strain energy prior to reaching peak stress. Furthermore, higher peak stresses are transmitted to deeper sections of the rock, thereby enhancing the overall pressure relief effect. Additionally, rock creep becomes more pronounced with greater pressure relief. The study also reveals that the degree of initial rock mass damage and the frequency of acoustic emission events increase with the efficiency of pressure relief. The findings suggest that an effective strategy to optimise borehole pressure relief involves first prioritising joint pressure relief through multiple boreholes, followed by the strategic expansion of borehole diameters, and ultimately, the increase in borehole depth.

Same calls, different meanings: Acoustic communication of Holocentridae.

The literature on sound production behaviours in fish in the wild is quite sparse. In several taxa, associations between different sound types and given behaviours have been reported. In the Holocentridae, past nomenclature of the different sound types (knocks, growls, grunts, staccatos and thumps) has been confusing because it relies on the use of several terms that are not always based on fine descriptions. Our study aims to ascertain whether holocentrids can produce a variety of sounds in the wild and if these sounds are associated with specific behaviours. Additionally, we aim to determine whether sounds produced by hand-held specimens, a common methodology to record sounds in standardised conditions in fishes, could correspond to some sounds produced by free-swimming individuals in natural conditions. Our study shows that all holocentrid species are able to produce sounds in 6 behavioural contexts of both agonistic (conspecific and heterospecific chases, competition) and social signalling types (acceleration, broadcasting, body quivering), in addition to previously described mobbing towards moray eels and symbiotic interactions with cleaner wrasses. In holocentrids, acoustic communication is not only based on single calls but can also involve series of sounds of different types that are arranged randomly. The large amount of combinations within acoustical events for each behaviour, resulting from both the quantity of sounds and their diversity, supports the absence of stereotypy. This suggests that sounds are produced to reinforce visual communication during the day in this family. Our results also suggest that sounds recorded by hand-held fishes are produced naturally in the wild. Our study challenges past nomenclatures and demonstrates sound critical function in augmenting visual communication, advancing our comprehension of acoustic ecology in teleost species.

Acoustic event detection for drone search and rescue system based on bi-directional long and short-term memory beamforming method to remove rotor noise

In outdoor search and rescue operations, drone carrying microphone array to collect acoustic signals has gradually been applied to search and rescue missions for people in distress. However, since the distance between the array and the drone is much smaller than the distance between the array to the trapped person, the drone rotor noise contaminates the collected acoustic signal seriously, which often makes the search and rescue system ineffective. It is necessary to study a method that can eliminate the influence of drone rotor noise on the detection results. To address this problem, an improved Minimum Variance Distortionless Response (MVDR) algorithm based on bidirectional long short-term memory network (BLSTM-MVDR) is proposed in this paper. This method includes the bidirectional long short-term memory network to estimate the time-frequency mask of drone rotor noise, which is used as the noise covariance matrix in the MVDR algorithm, to obtain the weight vector of the algorithm for eliminating the drone rotor noise, which removes the drone rotor noise component and improves the ability of the existing algorithms to resist the drone rotor noise. Finally, a Deep Learning-based sound event detection classifier is constructed by combining convolutional neural network to realize accurate and effective search and rescue of trapped people. To verify the accuracy and effectiveness of the proposed method, a complete drone search and rescue system is constructed using the server side of the hardware device and the client side of the software platform, and the effectiveness of the system in detecting the human voice event of trapped individuals is evaluated according to different signal-to-noise ratios, sound source directions and sound source distances. The results of this paper show that the effect using the method proposed in this paper is satisfactory, with the detection accuracy improved by 43.75 % and the effective search and rescue range doubled compared with the existing methods. This method improves the accuracy of outdoor search and rescue missions in a strong drone rotor noise interference environment, and promotes the development of drone technology applications.

Study on mechanical properties and acoustic emission characteristics of composite rock mass with different thicknesses of weak interlayer

To study the influence of the thickness of a weak interlayer on the deformation and failure characteristics of composite rock mass. Based on the measured data of sand-mudstone interbedded floor rock mass in the Yima mining area, Henan Province, China. Using quartz sand, gypsum, and cement as similar materials make composite rock-like specimens with weak interlayers. The mechanical properties, deformation evolution process, failure characteristics, and acoustic emission laws of rock samples with different interlayer thicknesses under uniaxial compression were studied using the digital image correlation method and acoustic emission monitoring technology. The results show that: (1) With an increase in the thickness of the weak interlayer, there is a corresponding decrease in the compressive strength of the specimen. When the thickness of the weak interlayer exceeds 5 cm, the compressive strength of the composite rock specimen is even less than that of the single weak rock-like. (2) Under uniaxial compression conditions, the strain concentration zone first appears in the weak interlayer part of the composite rock-like specimen. As the pressure continues to increase, the specimen is the first to fail at the position of the maximum strain concentration zone. (3) The thickness of the weak interlayer is an important factor affecting the failure mode of composite rock-like specimens. With the increase of the thickness of the weak interlayer, the composite rock-like specimens change from overall failure to local failure. (4) With the increase of the thickness of the weak interlayer in the middle of the composite rock specimen, the maximum value of the energy released by the single acoustic emission event and the maximum value of the cumulative acoustic emission energy decrease during the compression failure process. The research results can provide a reference for the manufacture of composite rock mass with a weak interlayer and the deformation and failure analysis of composite rock mass.

Comparison of the Sensitivity of Various Fibers in Distributed Acoustic Sensing

Standard single-mode telecommunication optical fiber is still one of the most popular in distributed acoustic sensing. Understanding the acoustic, mechanical and optical features of various fibers available currently can lead to a better optimization of distributed acoustic sensors, cost reduction and adaptation for specific needs. In this paper, a study of the performances of seven fibers with different coatings and production methods in a distributed acoustic sensor setup is presented. The main results include the amplitude–frequency characteristic for each of the investigated fibers in the range of acoustic frequencies from 100 to 7000 Hz. A single-mode fiber fabricated using the modified chemical vapor deposition technique together with a polyimide coating has shown the best sensitivity to acoustic events in the investigated range of frequencies. All of this allows us to both compare the studied specialty fibers with the standard single-mode fiber and choose the most suitable fiber for a specific application, providing an enhancement for the performance of distributed acoustic sensors and better adaptation for the newly aroused potential applications.

Hierarchical-Concatenate Fusion TDNN for sound event classification.

Semantic feature combination/parsing issue is one of the key problems in sound event classification for acoustic scene analysis, environmental sound monitoring, and urban soundscape analysis. The input audio signal in the acoustic scene classification is composed of multiple acoustic events, which usually leads to low recognition rate in complex environments. To address this issue, this paper proposes the Hierarchical-Concatenate Fusion(HCF)-TDNN model by adding HCF Module to ECAPA-TDNN model for sound event classification. In the HCF module, firstly, the audio signal is converted into two-dimensional time-frequency features for segmentation. Then, the segmented features are convolved one by one for improving the small receptive field in perceiving details. Finally, after the convolution is completed, the two adjacent parts are combined before proceeding with the next convolution for enlarging the receptive field in capturing large targets. Therefore, the improved model further enhances the scalability by emphasizing channel attention and efficient propagation and aggregation of feature information. The proposed model is trained and validated on the Urbansound8K dataset. The experimental results show that the proposed model can achieve the best classification accuracy of 95.83%, which is an approximate improvement of 5% (relatively) over the ECAPA-TDNN model.

Experimental Investigation on Rock Failure Characteristics of Large-Span Goafs Using Digital Image Correlation Analysis and Acoustic Emission Monitoring

Rockmass in deep mining is highly susceptible to large-scale collapses under high stress and blast-induced disturbances, leading to casualties and economic losses. To investigate the evolution characteristics of goaf instability and the types of seismic sources that induce instability, an experiment on goaf instability was designed under uniaxial compression conditions based on actual mining operations. The entire experimental process was monitored using digital image correlation analysis and acoustic emission monitoring. By calculating the digital speckle field on the surface of the rock specimen during the experiment, the evolution characteristics of the deformation and strain fields from the beginning of loading to complete failure were analyzed. The study explored the dynamic behavior of cracks from initiation to propagation and eventually inducing large-scale collapse. The results show that the instability process of the goaf begins with the formation of tensile cracks. As stress increases, shear cracks occur in the specimen, leading to macroscopic failure. Furthermore, based on the differences in overall microfracture types measured by RA-AF characteristic parameters during specimen failure, large amplitude acoustic emission events corresponding to the formation of dominant macroscopic cracks were selected, and the focal mechanisms of these events were inverted. The results indicate that shear failure sources are significantly more prevalent than tensile failure sources in acoustic emission events leading to goaf instability. These findings can provide useful guidance for the support design and the prevention and control of rockmass instability disasters.

Experimental study of type-I crack propagation in rock monitored by fiber Bragg grating

The occurrence of rock mass engineering disasters can be reduced significantly by obtaining the key information in the process of rock internal fracture accumulation and evolution through advanced monitoring means and taking certain preventive measures. Through fiber Bragg grating monitoring tests of the type-I crack propagation in rock, the spectral characteristics of the fiber Bragg grating in the process of rock failure were analyzed, the bandwidth expansion was extracted to characterize the local tensile strain during the crack propagation process, and the damage variable based on the bandwidth expansion was calculated to describe the type-I crack propagation process. The research results showed the following: (1) The fiber Bragg grating generated a non-uniform local tensile strain during type-I crack propagation. The bandwidth expansion of the wave spectrum was positively correlated with the local tensile strain. The change in the characteristics of the wave spectrum could provide the local real strain in the process of crack propagation. (2) During the steady propagation of the type-I crack, the bandwidth increased obviously, while the number of acoustic emission events was relatively small, and most of them were in the process zone. The gentle increase in the bandwidth corresponded to the unstable growth stage of the type-I crack, which could be used as a precursor of rock sample instability. During loading, the maximum tensile strain of the type-I crack was 6 %, while the maximum shear strain reached 1.8 %, primarily attributed to tensile failure. The average value of fracture toughness KIC, based on the equivalent linear elastic fracture mechanics model, was determined to be 1.21 MPa·m1/2, and the average value of fracture energy Gf was calculated as 57.7 N/m. The variation of peak reflectivity in fiber Bragg grating was associated with the closure of pre-existing defects and the propagation of microcracks in rocks, resulting in unstable shear strain at crack extension sites. (3) The damage variable D defined by the bandwidth broadening began to become concentrated and accumulated in the crack steady growth stage, and it increased slowly in the crack intensification and evolution stage. The grating was sensitive to changes when it was pasted to the surface at 90°, but it could easily fracture or become unable to capture spectral signals in time. Based on a comprehensive comparison, the monitoring effect was the best when the grating was pasted at 60° relative to the crack.

Automatic detection for bioacoustic research: a practical guide from and for biologists and computer scientists.

Recent years have seen a dramatic rise in the use of passive acoustic monitoring (PAM) for biological and ecological applications, and a corresponding increase in the volume of data generated. However, data sets are often becoming so sizable that analysing them manually is increasingly burdensome and unrealistic. Fortunately, we have also seen a corresponding rise in computing power and the capability of machine learning algorithms, which offer the possibility of performing some of the analysis required for PAM automatically. Nonetheless, the field of automatic detection of acoustic events is still in its infancy in biology and ecology. In this review, we examine the trends in bioacoustic PAM applications, and their implications for the burgeoning amount of data that needs to be analysed. We explore the different methods of machine learning and other tools for scanning, analysing, and extracting acoustic events automatically from large volumes of recordings. We then provide a step-by-step practical guide for using automatic detection in bioacoustics. One of the biggest challenges for the greater use of automatic detection in bioacoustics is that there is often a gulf in expertise between the biological sciences and the field of machine learning and computer science. Therefore, this review first presents an overview of the requirements for automatic detection in bioacoustics, intended to familiarise those from a computer science background with the needs of the bioacoustics community, followed by an introduction to the key elements of machine learning and artificial intelligence that a biologist needs to understand to incorporate automatic detection into their research. We then provide a practical guide to building an automatic detection pipeline for bioacoustic data, and conclude with a discussion of possible future directions in this field.

Characteristics of Deformation and Damage and Acoustic Properties of Sandstone in Circular Tunnel Morphology under Varying Inundation Depths

When water damage occurs in a mine, variations in the immersion levels of tunnels at different burial depths can be observed. There is a significant relationship between the stability of the surrounding rock and the depth of immersion. Therefore, studying the deformation and damage characteristics of sandstone with circular holes at varying immersion depths, along with their acoustic properties, plays a crucial role in maintaining the stability of water-rich roadways. The TAW-2000 press and static strain system were utilized to investigate the mechanical properties, crack evolution, and deformation field distribution of sandstone with circular holes at varying immersion depths. Additionally, this study analyzed the impact of immersion depth on the characteristic parameters of acoustic emission. The results indicate that immersion depth is negatively correlated with the compressive strength and modulus of elasticity of sandstone; as immersion depth increases, the duration of the compression and yield phases of the rock samples also increases, while the duration of the elastic phase remains relatively unaffected. Furthermore, greater immersion depths correspond to a decrease in the total number of cracks, although the proportion of tensile cracks increases, making the formation of secondary cracks less likely. The frequency of acoustic emission events (transient elastic waves generated by the formation, extension, or closure of tiny cracks within the rock) shows a closely correlated dynamic with the stress–time curve of the rock sample. The acoustic emission ringing counts generated by rock samples under submerged water conditions tend to stabilize with a slight increase before signs of rupture appear. Additionally, the cumulative total energy of acoustic emissions from the rock samples decreases as the water level rises. These research findings provide significant reference value for addressing issues related to water immersion and the extent of water saturation in roadways within rock engineering.

Detection of a stress related acoustic signature by passive acoustic monitoring in Atlantic salmon farming

Crowding is a necessary operation for moving fish out of or between sea cages in Atlantic salmon farming. These operations consist of gradually reducing available volume in the nets to increase the biomass density. This is known to stress the fish, increase risk of injury, and can in extreme cases increase mortality rates. To reduce the negative impacts of crowding, the operation must be adjusted based on real-time monitoring of fish welfare. Today, this is mainly done by manually monitoring surface activity, as the high fish density limits the use of available tools. In this paper a new approach for monitoring was applied to crowding of meso-scale cages containing 40 fish each, using automatic detection of passive acoustic signatures. Four synchronised hydrophones were deployed and the acoustic signature was defined from the recordings. The automatic detection algorithm recorded an extreme increase in registered acoustic events during crowding, as high as 1000 events per 10 min period, compared to only 20 events per 10 min period on the day following crowding. Correlating the number of acoustic events to recorded heart rate and activity measures from implants in the salmon indicate that automatically registering acoustic events is a promising method for monitoring farmed fish.

Enhancing Acoustic Data Mobilization: A User-Friendly Data Platform for the Asian Soundscape Monitoring Network

Passive Acoustic Monitoring (PAM) has emerged as a crucial tool in monitoring efforts to track environmental changes and evaluate conservation measures in response to the biodiversity crisis (Sugai et al. 2018). PAM now offers long-term and continuous insights into biodiversity, using acoustic indices correlating with biodiversity and deep-learning tools to aid in the detection and identification of animal sounds (Sueur et al. 2014, Kahl et al. 2021, Wu et al. 2022). However, managing and mobilizing original recordings remains challenging due to their large size and lack of structure, hindering broader applications of these valuable data. Established in 2014 by Academia Sinica and Taiwan Forestry Research Institute, the Asian Soundscape Monitoring Network has amassed over 20 million minutes of audible and ultrasonic recordings from diverse landscapes, including forests, wetlands, urban parks and farmlands across Malaysia, Thailand, Taiwan and Vietnam. To enhance data mobilization and utilization, we developed a user-friendly data platform enabling browsing, searching, visualization, exploration and retrieval of the original recordings (Fig. 1), as well as derived data such as acoustic indices and species occurrences, and associated weather records (Fig. 2) The platform also allows users to visualize the temporal dynamics of the soundscape and identify acoustic events at monitoring sites by examining long-term spectrograms and time series of acoustic indices (Fig. 3). Detailed information about the monitoring sites and recorder deployment is provided. Each recording is tagged with a Creative Commons license, and a unique Archival Resource Key is assigned to every data retrieval for persistent identification, facilitating data reuse. Additionally, recordings containing human voices are identified and restricted to protect privacy. Our aim with this data platform is to streamline the mobilization of acoustic data, foster diverse applications, and enhance the overall value of the data.

From Labquakes to Megathrusts: Scaling Deep Learning Based Pickers Over 15 Orders of Magnitude

AbstractThe application of machine learning techniques in seismology has greatly advanced seismological analysis, especially for earthquake detection and seismic phase picking. However, machine learning approaches still face challenges in generalizing to data sets that differ from their original training setting. Previous studies focused on retraining or transfer‐learning models for these scenarios, but require high‐quality labeled data sets. This paper demonstrates a new approach for augmenting already trained models without the need for additional training data. We propose four strategies—rescaling, model aggregation, shifting, and filtering—to enhance the performance of pre‐trained models on out‐of‐distribution data sets. We further devise various methodologies to ensemble the individual predictions from these strategies to obtain a final unified prediction result featuring prediction robustness and detection sensitivity. We develop an open‐source Python module quakephase that implements these methods and can flexibly process input continuous seismic data of any sampling rate. With quakephase and pre‐trained ML models from SeisBench, we perform systematic benchmark tests on data recorded by different types of instruments, ranging from acoustic emission sensors to distributed acoustic sensing, and collected at different scales, spanning from laboratory acoustic emission events to major tectonic earthquakes. Our tests highlight that rescaling is essential for dealing with small‐magnitude seismic events recorded at high sampling rates as well as larger magnitude events having long coda and remote events with long wave trains. Our results demonstrate that the proposed methods are effective in augmenting pre‐trained models for out‐of‐distribution data sets, especially in scenarios with limited labeled data for transfer learning.

An experimental setup to investigate relevant validation parameters for the auralization of commercial aircraft flyovers in complex urban contexts

Auralization has a key role in the design of immersive virtual reality tools for the study of human environments and their evolution. In soundscape and multisensory research, the validation of auralization frameworks traditionally relies on the impressions engendered by the virtual acoustic environment on the human subjects, and whether those impressions match real-world experiences. It uses subjective surveys related to the perceptual and affective spheres of individuals. This may lead to unclear validation. In this paper, we present an experimental setup to search for a set of relevant validation parameters for the auralization of complex urban environments involving commercial aircraft flyovers. Our investigation is tailored to the study of working and learning activities in offices or universities neighboring the flight paths. The experimental setup focuses on the quantitative evaluation of cognitive performance instead of the classic perceptual or affective personal survey-based approach. In the experiment, subjects undergo distinct acoustic environments, synthesized from real-world binaural recordings. Simultaneously, subjects perform memory, fluency, or puzzle-like tasks. Different cognitive performance parameters are measured throughout the experiment. The parameters most strongly correlated to the environments' acoustic events and properties are considered meaningful validation parameters.

Dynamic localization of unwanted events with machine learning-based method

In environmental acoustics measurements it is necessary to separate the contribution of the investigated source from the residual noise, by means of ambient noise measurements. Source separation is a complex task, especially for wind turbine noise, where long-term measurement campaigns are required. Although data analysis procedures that provide for this task are already in use, all of them require a preliminary step for anomalous events removal, i.e. events which are not characteristic of standard residual noise. This step can be complex and very time-expensive depending on the duration of the measurements and the measurement site. Depending on the specific site and source under evaluation, the variety of sounds that can be considered as unwanted for compliance with environmental acoustic regulations is heterogeneous, ranging from road vehicles, trains and planes to human and animal sounds and more. Nowadays, it is possible to extrapolate from an audio source the features that describe its semantics using machine learning. The approach presented in this paper aims to localize in time a specific event within a measurement containing several acoustic events, in order to label it and then proceed to remove the spurious events from the measurement.

Creating a specialized dataset for enhanced noise compliance in live concert environments

Regulatory guidelines mandate strict noise levels during concerts, specially for night leisure soundscapes, yet accurate compliance assessment is hindered by the presence of extraneous noises. This paper presents a methodology focused on creating a specialized dataset tailored to address this challenge. We have captured several live concerts, ensuring the dataset's relevance and applicability in noise compliance assessments. This paper exposes the methodology followed since the deployment of the sensors until the labeling of the most relevant parts of it by using the LAeq levels to select the most prominent sound sources. The dataset has been used to use a PANN and incorporate a classification mechanism into a noise monitoring platform for live concerts. By leveraging LAeq levels to identify prominent sound sources, our methodology ensures the inclusion of key acoustic events in the dataset, facilitating subsequent analysis and classification tasks. The integration of a PANN (Pretrained Audio Neural Network) enables the development of a robust classification mechanism tailored to the intricacies of live concert environments. This work underscores the importance of specialized datasets in addressing the challenges of noise compliance in concert settings, offering a modest resource for advancing research and practical applications in this field.

Fish feeding behavior quantification and automatic acoustic event detection: algorithmic optimization with deep learning

In recent years, the integration of sensor processing technologies with intelligent feeding control systems has become crucial for enhancing recirculating aquaculture systems (RAS). Intelligent feeding control significantly improves profitability and animal welfare in aquaculture by accurately monitoring feed intake and reducing waste. Acoustic methods have proven effective in automating feed monitoring, particularly in assessing feeding intensity. Automatic Audio Event Detection (AED) holds promise for identifying sound events associated with feeding activities, enabling precise monitoring of feeding behavior and estimation of feed consumption through frequency-based feeding signatures. Advancements in sensor technologies and computational capabilities enable real-time collection, processing, and analysis of acoustic data. However, handling high volumes of data with minimal latency and high accuracy poses challenges. To address these challenges, we propose a systematic approach: (1) Data preprocessing using wavelet transforms enhances efficiency and reduces computational overhead. (2) Data pruning focuses on informative features and reduces input data dimensionality, improving classification accuracy. (3) We adopt a weakly-labeled semi-supervised learning model with Convolutional Neural Networks (CNNs) to detect feeding events, improving classification performance. (4) Ensemble learning combines multiple classifiers to mitigate overfitting and enhance model generalization. Our methodology shows promise in accurately quantifying fish feeding behavior using acoustic signals.

A signal processing system for infrasound-based detection and tracking of tornadoes

It has been established that tornadic storms emit infrasound at infrasonic frequencies and that this sound can be detected from practically significant distances exceeding 100 km in some cases. This work describes a signal processing system that is used to ingest data from a network of infrasound arrays and produce near real-time detections and ground tracks of multiple, simultaneously active tornadic storms. Specifically, a brief review of a particular array network and corresponding hardware is presented followed by detailed descriptions of the array signal processing architecture and source geolocation system. Emphasis is placed on algorithms for multiple moving source acoustic source bearing tracking, low SNR signal detection, mitigation of errors created by anomalous acoustic events, and multiple-array data association. Performance results corresponding to simulated and recorded data are presented.