In the digital era of increasing software complexity, improving the development efficiency of safety-critical software is a challenging task faced by academia and industry in domains such as nuclear energy, aviation, the automotive industry, and rail transportation. Recently, people have been excited about using pre-trained large language models (LLMs) such as ChatGPT and GPT-4 to generate code. Professionals in the safety-critical software field are intrigued by the code generation capabilities of LLMs. However, there is currently a lack of systematic case studies in this area. Aiming at the need for automated code generation in safety-critical domains such as nuclear energy and the automotive industry, this paper conducts a case study on generating safety-critical software code using GPT-4 as the tool. Practical engineering cases from the industrial domain are employed. We explore different approaches, including code generation based on overall requirements, specific requirements, and augmented prompts. We propose a novel prompt engineering method called Prompt-FDC that integrates basic functional requirements, domain feature generalization, and domain constraints. This method improves code completeness from achieving 30% functions to 100% functions, increases the code comment rate to 26.3%, and yields better results in terms of code compliance, readability, and maintainability. The code generation approach based on LLMs also introduces a new software development process and V-model lifecycle for safety-critical software. Through systematic case studies, we demonstrate that, with appropriate prompt methods, LLMs can auto-generate safety-critical software code that meets practical engineering application requirements. It is foreseeable that LLMs can be applied to various engineering domains to improve software safety and development efficiency.

Objectives: The purpose of this study was to objectively quantify and evaluate the ergonomic risk of clinical physiotherapy practices and evaluate physiotherapists for work-related musculoskeletal disorders and pain.Methods: Twenty-nine physiotherapists in the rehabilitation department of a large-scale tertiary hospital were recruited in this study. The sampling period lasted for 2 weeks for each physiotherapist and interval sampling was adopted to avoid duplication of cases. Therapist posture during physiotherapy was captured, tracked and analyzed in real time using structured light sensors with an automated assessment program. The quantification of ergonomic risk was based on REBA (Rapid Entire Body Assessment) and the RPE (perceived physical exertion) scores of the therapists were recorded before and after treatment, respectively.Results: Two hundred and twenty-four clinical physiotherapy cases were recorded, of which 49.6% were high risk and 33% were very high risk, with none of the cases presenting negligible risk. The positioning (p < 0.001) of physiotherapist had a considerable impact on ergonomic risk and pediatric physiotherapy presented a higher risk to physiotherapists than adults (p < 0.001). The RPE score of physiotherapist after performing physiotherapy was greater than before physiotherapy and was positively correlated with the REBA distribution.Conclusion: Our study creates an automatic tool to assess the ergonomic risk of physiotherapy practices and demonstrates unacceptable ergonomic risk in common practices. The high prevalence of musculoskeletal disorders and pains recommends that rehabilitation assistance devices should be optimized and standard ergonomic courses should be included in physiotherapists' training plans.

Background: This study aims to investigate the associations between signal-level physical activity (PA) features derived from wrist accelerometry data and cognitive status in older adults, and to evaluate their potential predictive value when combined with demographics. Methods: We analyzed PA data from 3,363 older adults (NHATS: n = 747; NHANES: n = 2,616), with each participant contributing a complete 3-day continuous activity sequence. We extracted the most relevant PA features associated with cognitive function using feature engineering and recursive feature elimination. Demographic characteristics were also incorporated, and multimodal data fusion was achieved through canonical correlation analysis. We then developed explainable machine learning models, primarily random forest, optimized with hyperparameters, to predict individual cognitive function status. Results: Using recursive feature elimination, we identified the top 20 PA features from each dataset and combined them with demographic features for modeling. The models achieved AUCs of 0.84 and 0.80 for NHATS and NHANES. Change quantiles and FFT coefficients emerged as the consistently top-ranked PA features across datasets, ranking 1st and 2nd respectively in their predictive importance for cognitive function. Models based on the top 10 PA features common to both datasets, along with demographic features, achieved AUCs above 0.8. Conclusions: This study identifies novel time-frequency domain features of physical activity that show robust associations with cognitive status across two independent cohorts. These features, particularly those capturing activity variability and rhythmicity, provide complementary information beyond traditional cumulative PA measures. Based on these findings, we developed a proof-of-concept application that demonstrates the feasibility of translating these PA analytics into practical monitoring tools in real-world settings.

Epilepsy affects over 70 million individuals worldwide, with optimal physical activity (PA) levels remaining challenging to determine due to potential negative outcomes from both insufficient and excessive activity. To quantify the associations between objective PA and mortality in people with epilepsy, we analyzed accelerometer data from 98,561 UK Biobank participants, including 1167 with epilepsy, to quantify associations between objectively measured PA and mortality. During a median follow-up of 7.1 years, people with epilepsy had significantly higher mortality rate (Standardized mortality ratio: 2.39, 1.97–2.86). Higher sedentary behavior duration was associated with lower all-cause mortality (hazard ratio: 0.87, 0.78–0.97). Dose-response analyses identified sedentary durations of 7–13 h/day (p nonlinear=0.025) and moderate-to-vigorous PA of 0.1–0.2 h/day (p nonlinear=0.005) were associated with lower mortality. An explainable machine-learning model that combined multi-dimensional PA with demographic and health information effectively stratified individual risk (Area Under the Receiver Operating Characteristic curve: 0.87 ± 0.08) and could support personalized activity guidance through a wearable system.