Abstract To address the challenge of decommissioning transmission towers, this study proposes a digital twin-enhanced Multi-Agent Reinforcement Learning (MARL) framework to overcome the low efficiency, high risk, and poor adaptability of manual and existing robotic demolition methods. The core innovation lies in transforming the digital twin from a passive monitoring tool into an active predictive engine within the MARL decision loop. This enables high-fidelity state mapping and forward-looking risk simulation, thereby enhancing collaborative decision-making capabilities in dynamic environments. The framework integrates four key technologies: 1) A high-precision hardware architecture based on EtherCAT to eliminate communication delays and ensure precise motion trajectories; 2) An enhanced kinematic model with real-time thermal compensation to suppress positioning drift caused by temperature variations; 3) A digital twin mechanism integrating multi-scale mechanical models with real-time sensing, providing high-precision state mapping and risk rehearsal to support collaborative operations; 4) A multi-objective MARL decision engine leveraging the digital twin's precise state to optimize dynamic strategies, ensuring safety and coordination. Through high-fidelity simulation and real-time synchronization, the digital twin reduces MARL training time by 40% and accelerates strategy convergence by 2.3 times. Experiments demonstrate that compared to traditional manual and single-robot operations, the system achieves sub-millimeter trajectory accuracy, over 50% efficiency improvement, and over 60% reduction in accident rates, while maintaining high adaptability and coordination performance under adverse conditions. This research delivers a high-precision, low-risk, unmanned solution for intelligent tower demolition. The proposed framework is also applicable to high-risk precision operations such as bridge demolition and ship dismantling, establishing a new paradigm for the green decommissioning and maintenance of critical infrastructure.

Bullying in educational settings remains a critical issue, particularly for students with Autism Spectrum Disorder (ASD) who often struggle with verbal communication and social cues. This research addresses the lack of affordable and non-invasive early warning systems designed specifically for the sensory needs of autistic children. The primary objective was to develop and test an Arduino- and Internet of Things (IoT)-based early warning bell system to empower ASD students to report bullying incidents in real time. Using a Research and Development (R&D) approach with the 4D development model (Define, Design, Develop), the system was tested by 50 stakeholders, including special education teachers and experts. Results indicate a high level of system effectiveness, with a 94% rapid warning response rate and a 90% increase in students' perceived sense of safety. Statistical analysis yielded a Cronbach’s Alpha of 0.948, confirming high reliability. This study concludes that integrating simple IoT push-button technology with Telegram notifications provides a feasible, low-cost solution for inclusive schools. It shifts the safety paradigm from passive monitoring to active student empowerment, offering a sustainable protection strategy for vulnerable learners.

Abstract Understanding the acoustic communication of the endangered main Hawaiian Islands insular population of false killer whales (Pseudorca crassidens) is essential for effective management. In this study, biologging tags were deployed on four individuals, recording 26.2 h of acoustic data. A total of 5940 high-quality possible focal pulsed calls were analysed and 52 stereotyped call types were characterized. The fundamental frequency contour ranged from a mean minimum frequency of 7.22 ± 0.78 to 9.28 ± 0.80 kHz (mean maximum frequency) with a mean duration of 0.32 ± 0.08 s. Predominant call types and call rates across dive states varied by individual. Probability of calling was higher during the descent and bottom phase compared with the surface. Four types of nonlinear phenomena (NLP) were documented including biphonation (call + clicks, 78% of all NLP), secondary sidebands, chaos and frequency jumps. Frequency jumps were commonly produced by two animals from the same group (13–19%). Most calls contained NLP (80%). The average call rate recorded by the tags on two individuals was 17.5 calls m−1 (hourly average) and call rates decreased as swim speed increased. Our findings suggest high repertoire diversity and high signal complexity informing future passive acoustic monitoring efforts.

Biodiversity monitoring is crucial for understanding species trends and their responses to anthropogenic change. Passive acoustic monitoring (PAM) offers a scalable, non-invasive approach to capture ecological information across large spatial and temporal scales. However, it generates vast amounts of audio recordings, whose management and analysis present technical challenges. To support diverse user needs in ecoacoustic research, a growing number of software tools have emerged, but the landscape remains fragmented and difficult to navigate. We provide a systematic overview of software tools used for soundscape assessment across terrestrial, freshwater, and marine environments. We screened peer-reviewed literature and complemented it with database cross-checking to identify and categorize tools according to four PAM data workflow components: data management, signal pre-processing, visualisation and navigation, and acoustic analysis. We found 221 available tools of which 174 were explicitly designed for PAM. Most tools were freely accessible (83%) with only a smaller fraction being commercial (12%) or limited access (5%). Terrestrial research accounted for most software mentions (476 studies), followed by aquatic (319) and cross-realm (64) studies. Nearly half (45%) were package-based frameworks within R, Python, or MATLAB. Acoustic analysis was the most represented workflow component, while only 40 tools covered all four of them. This diversity illustrates the field’s rapid technical growth but also its redundancy and methodological fragmentation: to date, many tools target only a subset of workflow components and replicate similar functionalities. Despite this, the prevalence of PAM-dedicated software indicates increasing specialization and technical maturity within ecoacoustics. Our structured inventory underscores the need for greater collaboration and continuity in software development, promoting the improvement and accessibility of existing tools rather than further proliferation. This living systematic review, provides a practical, biannually updated reference for tool selection and fosters transparency, comparability, and cooperation across bioacoustic and ecoacoustic research communities.

Abstract. Wind forcing plays a pivotal role in driving upper-ocean physical and biogeochemical processes, yet direct wind observations remain sparse in many regions of the global ocean. While passive acoustics have been used to estimate wind speed from moored and mobile platforms, their application to profiling floats has been demonstrated only in limited cases. Here we report the first deployment of a biogeochemical profiling float equipped with a passive acoustic sensor explicitly designed for wind retrieval, aimed at detecting wind-driven surface signals from depth. The float was deployed in the northwestern Mediterranean Sea near the DYFAMED (DYnamique des Flux Atmosphériques en MEDiterranée) meteorological buoy from February to April 2025 and operated at parking depths of 500–1000 m. We demonstrate that wind speed can be successfully retrieved from subsurface ambient noise using established acoustic algorithms, with float-derived estimates showing good agreement with collocated surface observations. To evaluate scalability to remote regions, we simulate a remote deployment scenario by refitting the acoustic model of Nystuen et al. (2015) using ERA5 reanalysis as a reference for surface wind. The ERA5-based calibration performs well under moderate winds but exhibits systematic high-wind bias (≥ 10 m s−1). Finally, we apply a residual learning framework to correct these estimates using a limited subset of DYFAMED wind data, simulating conditions where only brief surface observations are available. The corrected wind time series achieved a 38.6 % reduction in RMSE, demonstrating the effectiveness of combining reanalysis with sparse in-situ calibration. This framework improves agreement with in-situ wind observations relative to reanalysis alone, supporting a scalable strategy for float-based wind monitoring in data-sparse ocean regions. Such capability has direct implications for improving estimates of air–sea exchanges, interpreting biogeochemical fluxes, and advancing climate-relevant ocean observing.

As the aging population grows and more elderly individuals live independently, the demand for reliable, unobtrusive home health monitoring becomes increasingly important. Existing in-home health monitoring systems often face limitations such as privacy concerns, dependence on unreliable wearable devices, degraded accuracy in complex environments, and lack of continuous monitoring capability. To address these challenges, we propose a long-term home health monitoring system that primarily relies on audio sensing, supplemented by other noninvasive modalities. Our approach is able to accurately detect and recognize overlapping acoustic events with fine-grained temporal resolution, surpassing conventional audio-based methods for activity recognition. The system incorporates a transformer-based time-frequency fusion module and a category dynamic threshold strategy to improve detection performance under semi supervised conditions. Experiments on real-world dataset demonstrate that our method outperforms existing baselines, achieving PSDS$_{1}$, PSDS$_{2}$, and EB-F1 scores of 0.581, 0.930 and 55.1%, with improvements of 0.054, 0.019, and 2.3%, respectively. In addition, a 30 day field deployment involving 10 elderly participants confirms the robustness and practicality of the system for real-world applications. By allowing continuous passive monitoring of daily activities and abnormal acoustic events, our system has significant potentials for early detection of health risks, behavioral anomalies, and long-term wellness tracking in aging in place scenarios.

BackgroundPassive digital monitoring of real‐world behaviors holds potential for identifying cognitive decline in Alzheimer's disease and related dementias (ADRD). However, little is known about how privacy concerns may affect older adults’ participation in these studies, or how to educate older adults on mitigating privacy risks. This study aimed to characterize digital privacy concerns in older adults enrolling in a passive digital monitoring study and evaluate the efficacy of an educational intervention.MethodFifty‐one older adults (mean age: 74.94±5.29 years) enrolled in this study, with 32 classified as cognitively‐normal and 19 cognitively‐impaired (MoCA‐blind cut‐off=18). At baseline, participants answered questions addressing level of general and study‐specific digital data privacy concerns and self‐perceived knowledge on maintaining data privacy. Following a brief educational presentation reviewing data protection measures in this study (e.g., specifying data collected/not collected, instruction on deactivating collection of specific data elements), privacy concerns were re‐assessed. A 2 (Time; pre/post) x 2 (Condition; general/study‐specific) x 2 (Cognitive Status; cognitively‐normal/impaired) mixed model ANOVA analyzed the effect of the educational intervention on privacy concerns, and a 2 (Time) x 2 (Cognitive Status) mixed model ANOVA analyzed its effect on self‐perceived knowledge.ResultIn the privacy concern model, there was a significant within‐subjects effect of Condition, F(1,49)=13.69, p <.001, ηp2=.22, with greater concern about general versus study‐specific data privacy. A significant Condition x Cognitive Status interaction, F(1,49)=4.16, p = .05, ηp2=.08, indicated cognitively‐impaired participants endorsed lower concern about general privacy than cognitively‐normal participants. A significant main effect of Time, F(1,49)=23.62, p <.001, ηp2= .33, indicated privacy concerns were reduced following the intervention. The remaining two‐ and three‐way interactions were not significant, p's >.21. No significant main effects or interactions were found in the knowledge model, p's >.09.ConclusionOlder adults with and without cognitive impairment expressed greater confidence in their study‐related digital data privacy despite having general privacy concerns. A brief educational intervention effectively reduced participants’ privacy concerns. Results suggest educational efforts by research teams can attenuate digital data privacy concerns in older adults participating in ADRD digital monitoring studies. As digital tools are increasingly incorporated in ADRD studies, it is crucial researchers prioritize participants’ digital data privacy.

Scaling up bedload monitoring: a passive acoustic approach for large river systems.

Context Species monitoring programs are essential for assessing population status and informing management decisions. N-mixture models have become a widely used tool for estimating population growth rates (λ) from passive monitoring data, such as counts from camera traps or sonar surveys, because they account for imperfect detection and are applicable to common data types in fisheries and wildlife studies. Aims Our objective was to evaluate how site selection strategies and the proportion of sites sampled influence the accuracy and precision of λ estimates in single-season and dynamic N-mixture models. We were particularly interested in whether preferential sampling—the common practice of focusing monitoring efforts on sites known or presumed to have higher densities—could bias estimates of λ when using these modeling approaches. Methods We simulated count data from populations with spatially varying growth rates that led to increasing, stable, or decreasing overall growth, under scenarios of high or low detection probability. These data were analyzed using single-season and dynamic N-mixture models, with different sampling designs, including random and preferential site selection, and varying the proportion of sites sampled. Key Results Precision of λ estimates was similar between single-season and dynamic models, though dynamic models offered potential gains by borrowing information across years. However, preferential site selection consistently introduced bias into λ estimates. We also found more precise values in declining populations, likely due to reduced dispersion from the mean as population sizes contract, compared to the greater variability observed in expanding populations. Conclusions Our findings demonstrate that preferential sampling can bias population growth estimates from N-mixture models, even when using sophisticated model structures. Random site selection across all potential monitoring locations consistently provided less biased estimates of λ, reinforcing the importance of sampling design. Implications Recognizing and accounting for potential sampling biases is critical when applying N-mixture models to assess species trends. Our results offer practical guidance for designing or evaluating monitoring programs, emphasizing the need for randomized site selection to ensure accurate and unbiased population growth estimates.

The number of species threatened with extinction is continuously increasing, underscoring the need for reliable population estimates to develop effective conservation plans. The ability to confirm a species' presence during surveys (i.e., detectability) is central for population estimates. While audio-visual sensors, like camera traps and passive acoustic monitoring (PAM), have emerged as valuable tools for monitoring primates, few studies have systematically compared their detectability, particularly in dense forests with limited visibility and for elusive species. Here, we compared 40-days continuous monitoring with audio-visual sensor (camera traps, N = 19; PAM, N = 7) versus human-based point transects with three survey visits (N = 20) on wild moor macaques (Macaca maura) in two different habitats: forest (N = 10) and open areas (N = 10). Using occupancy models to compare the detection probability (p), we found that camera traps (p = 0.63 ± 0.04) and PAM (p = 0.79 ± 0.08) outperformed point transects (p = 0.33 ± 0.07), regardless of habitat type. After equalizing survey time between methods, we found that detections were greater on point transects in surveys shorter than 1 day, but camera traps and PAM equalized their performance with two survey days (p-value < 0.05). Notably, combining both audio-visual sensors yielded the highest detectability (p = 0.87 ± 0.05). These results highlight the effectiveness of audio-visual sensors and support multi-method approaches for monitoring primates in tropical forests. Overall, this research contributes to designing more effective monitoring protocols for primate species, which are essential for planning conservation strategies.

Passive acoustic monitoring reveals the limited distribution of an indicator species, the White-headed Woodpecker ( Leuconotopicus albolarvatus ), in the northern Blue Mountains, USA

Suggestions for extending, expanding, and integrating passive acoustic monitoring programs

Passive acoustic monitoring predicts higher avian diversity metrics than traditional bird surveys across multiple Australian bioregions

Research on Passive Monitoring Mechanisms and Optimization for Oil-Water Interface in Salt Cavity Reservoirs Based on Geothermal-Driven Thermal Perturbations

This study explores the use of Passive Acoustic Monitoring (PAM) to assess the hunting pressure on the vulnerable Red-breasted Goose ( Branta ruficollis ) population wintering in Romania. The Red-breasted Goose is a long-distance migrant whose numbers have been declining, with south eastern Romania serving as a significant wintering area. Despite restrictions on hunting, the species remains vulnerable to accidental shooting and gunshot disturbances targeting Greater White-fronted Geese ( Anser albifrons ), a legal quarry species that coexists in mixed flocks with the Red-breasted Goose. This research employed AudioMoth acoustic devices at three Natura 2000 sites during two winter seasons (2018&amp;ndash;2020). The devices recorded over 5900 hours of sound data, detecting 2530 gunshots across 220 days. Results showed that the hunting pressure was highest at Balta Alb&amp;#259; Lake, where 69% of study days recorded gunshots. The findings highlight that PAM is a cost-effective method for monitoring hunting activity, with potential applications for enforcing conservation measures and preventing poaching. The study concludes with recommendations for improving field data collection, data analysis, and implementing stronger conservation measures, such as hunting bans within critical areas for the Red-breasted Goose.

ABSTRACT Passive acoustic monitoring (PAM) is a valuable tool for studying habitat use patterns of cetaceans, often broadening the spatial and temporal scope of traditional visual survey effort. For endangered populations, understanding habitat use is critical for effective conservation and management, as increasing anthropogenic pressures may alter residency and occurrence. Here, we analyse near-continuous, multi-year PAM data to assess summer presence and residency patterns of the endangered southern resident killer whale population (SRKW; Orcinus orca ater) in Canadian Pacific waters. Acoustic data were collected over a 6-month period (1 May–31 Oct) across 3 years (2018–2020) at Swiftsure Bank and in the Strait of Juan de Fuca, with recorders positioned off Port Renfrew, Jordan River and Sooke. Seasonal acoustic occurrence of SRKW was markedly higher at Swiftsure Bank and Port Renfrew compared to the other sites, with a pronounced peak in acoustic presence in August. Diel detection of SRKW differed among sites, with lower call detection at dawn and higher detections at dusk at Swiftsure Bank, but no discernible diel patterns at Sooke. These findings improve our understanding of SRKW summer habitat use patterns within part of their critical habitat and highlight methodological considerations for optimising PAM deployments and post-processing procedures.

Abstract This article explores how greater interplay between science, technology, and art can facilitate deeper connectedness with nature and large, distributed natural processes, such as global warming and avian migration. The author, an ecoacoustic data scientist and musician, lays a foundation by discussing how the sublime aesthetic characterizes our appreciation of nature, the myriad benefits of nature immersion, and the ability of art---particularly sound art---to facilitate a connection with natural hyperobjects. The author then applies these concepts in several sound art pieces derived from bird detections within the Sound of Norway, a national-scale passive acoustic monitoring network.

Abstract Biodiversity, particularly insects, faces considerable threats in intensively managed agricultural landscapes. Agroforestry systems (AFS), which integrate woody elements into agricultural land, can enhance biodiversity. This study aims to identify management factors influencing orthopteran richness and abundance in AFS. Additionally, it evaluates the utility of passive acoustic monitoring (PAM) for orthopteran detection by comparing it to field monitoring. Orthopteran monitoring was conducted at 20 silvoarable AFS in western Switzerland. Orthopteran richness was recorded using transects and PAM, while abundance was obtained only from transects. Both methods yielded similar species numbers. Daytime PAM detected cryptic or low-abundance species missed by daytime transects but failed to record one non-stridulating and some nocturnal species. Consequently, data from both methods were combined to provide a more comprehensive analysis of factors influencing orthopteran richness. The analysis revealed that increasing plant species diversity within the understory vegetation strips (UVS) had a positive, though not statistically significant, effect on orthopteran species richness. Furthermore, a non-significant reduction trend in orthopteran abundance was observed in conventionally managed agroforestry systems compared to organically managed ones. Implications for insect conservation Our study shows that PAM can effectively monitor orthopteran richness in AFS. By increasing plant diversity in UVS and through organic management, farmers can enhance orthopteran richness and abundance in AFS and support biodiversity conservation.

This study integrates data on the prevalence, infection dynamics and risks associated with African swine fever virus (ASFV) outbreaks in Croatian wild boar during 2023–2024. Although the overall ASFV DNA prevalence in Croatia was 0.24%, the highest prevalence (2.29% in 2023 and 4.69% in 2024) was recorded in Vukovar-Srijem County. Genetic typing identified ASFV genotype II, subgroup 19, consistent with strains isolated from domestic pigs in Croatia and circulating in neighboring countries. Anti-ASFV specific antibodies were detected in 10.34% of wild boar tested in counties with previously reported DNA findings. In Vukovar-Srijem County, 4.60% of wild boar were positive for both, ASFV DNA and antibodies, suggesting ongoing virus infection, whereas the proportion of boar positive only for antibodies was 5.75%, indicating survival of acute infection. Statistical analysis revealed an increase in ASFV DNA detection from 2023 to 2024 (p = 0.043), with a higher prevalence in carcasses than in hunted animals (p = 0.001), highlighting the need for passive monitoring. While gender showed no statistical significance, a higher infection rate was observed in older animals (p = 0.001). The identified course of infection involved spillover events between domestic pigs and wild boar, with a significant anthropogenic influence.

A wave glider for passive acoustic monitoring of cetaceans and anthropogenic sources in the central Mediterranean Sea