Reinforcement learning-based adaptive traffic signal controllers have traditionally prioritised efficiency, with only recent studies beginning to incorporate safety considerations into signal design. However, most safety-oriented approaches rely on simulated trajectories that oversimplify driver behaviour, leading to potentially unreliable safety estimates. In contrast, studies based on real-world traffic conflicts often depend on rare critical events and fail to capture the safety conditions across the broader range of traffic interactions. Moreover, existing safety-oriented controllers typically employ fixed state representations tied to specific intersection layouts, limiting the transferability of learned signal policies across different geometric configurations. To address these limitations, this study proposes a graph-based reinforcement learning framework to optimise signal timings for both safety and efficiency. The safety objective minimises time spent in conflict, measured using the proportion of stopping distance, while the efficiency objective reduces queue length. A multi-objective deep reinforcement learning algorithm was used to optimise both objectives simultaneously. The intersection was represented as a graph of lanes and movements, which was processed through a graph neural network to capture underlying traffic dynamics. Video data were collected over two weekdays from two four-legged signalised intersections in Queensland, Australia. Road user trajectories were extracted using artificial intelligence-based video analytics and subsequently used to calibrate a microscopic traffic simulation environment via genetic algorithms. The trained signal control policy was then evaluated within these calibrated environments and compared with the existing field signal plan, as well as safety-oriented and efficiency-oriented reinforcement learning variants. Relative to the field signal plan, the proposed safety–efficiency framework achieved reductions of 5.1 % in average waiting time, 7.9 % in queue length, and 17.3 % in time spent in conflict. When transferred to a new intersection with a different layout, the learned policy continued to outperform the field signal plan across both safety and efficiency metrics. These findings demonstrate that traffic signals can be redesigned to improve safety without compromising efficiency. Overall, the proposed graph-based reinforcement learning framework offers a promising direction for intelligent signal control systems.
Abstract Construction workers worldwide grapple with a psychosocial wellbeing crisis characterized by elevated rates of stress, depression, substance abuse, and suicide. Although research interest has grown, existing studies have predominantly relied on small, cross-sectional surveys that cannot capture how stakeholder concerns evolve over time or reflect the authentic voices of those most affected. This investigation employed machine learning techniques to analyze 2,368 social media comments from X, YouTube, and Reddit spanning 2015 to 2022 and revealed how construction communities discuss psychosocial wellbeing when speaking freely in digital spaces. Structural topic modeling identified five dominant themes: nature and context of work (33.8%), industry environment and culture (22.6%), mental health conditions and outcomes (16.6%), interventions (15.5%), and mental health literacy (11.5%). Temporal analysis revealed a significant shift in how these conversations unfolded. Early discussions centered heavily on problems such as toxic workplace cultures and adverse health outcomes, but these problem-focused themes declined significantly ( p < 0.01 ) as solution-oriented discourse gained momentum. By 2021, discussions about interventions had overtaken conversations about mental health conditions and outcomes, marking a decisive shift from crisis articulation to action-seeking. Network analysis revealed that interventions correlated positively with mental health literacy ( r = 0.638 ) while showing strong negative associations with problem-focused topics ( r = − 0.777 ), indicating that solution-seeking behavior actively displaces problem-centered discourse. These findings suggest that the industry’s psychosocial wellbeing discourse has reached a critical turning point, presenting opportunities to implement interventions aligned with demonstrated stakeholder readiness for change. This study contributes to the body of knowledge by providing the first longitudinal social media analysis of psychosocial wellbeing discourse in construction as a scalable complement to cross-sectional surveys, documenting a problem-to-solution discursive transition that signals readiness for change, and establishing quantifiable relational patterns among mental health themes to guide integrated intervention strategies.
Background Fear of cancer recurrence (FCR) is a top-rated concern and unmet need for support among gynaecological cancer patients yet seldom discussed during healthcare consultations. Clinicians report several challenges in initiating conversations about FCR with patients and limited practical guidance is available. Objective This study aimed to identify communication practices for initiating conversations about FCR during outpatient consultations after treatment for gynaecological cancer. Methods A collection of 30 audio-recorded outpatient consultations, involving 4 gynae-oncology clinicians, 30 people who had completed treatment for endometrial or ovarian cancer, and 11 accompanying family members and friends, were reviewed to identify instances where FCR was discussed. These instances were analysed using conversation analysis. Results Six of the 30 consultations, all involving the same gynae-oncology clinician, included talk about FCR. Two recurrent communication practices were used to initiate these conversations: (1) Universal claims (e.g., “I always ask…”) framed talk about FCR as a routine part of post-treatment follow-up for all gynaecological cancer patients, and (2) General claims (e.g., “Some people find that…”) normalised FCR post-treatment and provided opportunity for the patient to indicate if FCR was relevant to them and to describe their experience. Both practices provided avenues for the gynae-oncology clinician to discuss available support for FCR. Conclusions FCR was not regularly discussed in this collection of recordings, reflecting previous reports of clinical practice. Nonetheless, direct observation and detailed analysis of real-word consultations where FCR was discussed identified two simple strategies that clinicians can use to initiate these conversations. First, universal claims can shift the conversation to FCR specifically and frame this talk as routine. Second, general claims normalise FCR and invite patients to share their experiences. Future research could investigate use of these practices across multiple clinicians and patient groups and examine the impact of FCR conversations for patients and families.
Prolonged skin contact with wearable sensors can promote perspiration-induced bacterial accumulation and biofilm formation, potentially compromising device performance and long-term health monitoring. Here, we introduce a flexible 3D piezoelectric ZnO tetrapod–poly(vinylidene fluoride) (ZnO-Ts/PVDF) composite that enables on-demand antibacterial activity without relying on antimicrobial-agent leaching. The unique 3D morphology of ZnO tetrapods enhances charge generation under mechanical deformation and ultrasonic activation, promoting localized electron transfer and reactive oxygen species (ROS) formation that contribute to bacterial membrane damage. The composite exhibited > 98% bactericidal efficiency under mechanical stimulation and > 90% under ultrasonic stimulation. Wearable time-course studies on human skin at 8, 16, and 24 h demonstrated a substantially lower recoverable culturable bacterial burden on ZnO-Ts/PVDF than on pristine PVDF under both dry (forearm) and humid (toe) conditions. The film also exhibited excellent durability, retaining > 95% antibacterial activity over 20 days. Collectively, these findings demonstrate a flexible and reusable piezoelectric antibacterial platform in which morphology-assisted membrane disruption and stimulation-induced oxidative processes contribute to antibacterial activity. This platform provides a promising strategy for self-cleaning, skin-contact wearable and biomedical interfaces.
Although institutional quality is widely recognised as important for environmental performance, less is known about whether the governance–energy relationship varies across levels of institutional capacity. This study examines the relationship between inclusive governance, energy intensity, and CO2 emissions volatility across 71 developing economies in the context of an increasingly fragmented global energy system. We focus on whether the governance–energy relationship varies across levels of institutional capacity. Inclusive governance is measured using a composite index incorporating political institutional quality, gender representation, and income equality. The empirical analysis uses a piecewise panel specification based on the sample median of inclusive governance, with system-GMM and local-projection estimates providing supplementary sensitivity evidence. Inclusive governance is negatively and significantly associated with energy intensity in both governance regimes in the baseline specification, with a somewhat stronger association in the higher-governance regime. However, this nonlinear pattern is sensitive to alternative estimators and breakpoints and is therefore interpreted as suggestive rather than definitive. In contrast, inclusive governance is not significantly associated with CO2 emissions volatility across most specifications. These findings indicate that the relationship between governance and environmental performance may be outcome-specific: institutional capacity is more clearly associated with energy efficiency than with emissions volatility. The results are interpreted as conditional associations rather than causal effects and suggest that the environmental relevance of governance may vary with both institutional capacity and the outcome being examined.
