Precise Model Research Articles

A Coupled Model for Forecasting Spatiotemporal Variability of Regional Drought in the Mu Us Sandy Land Using a Meta-Heuristic Algorithm

Vegetation plays a vital role in terrestrial ecosystems, and droughts driven by rising temperatures pose significant threats to vegetation health. This study investigates the evolution of vegetation drought from 2010 to 2024 and introduces a deep-learning-based forecasting model for analyzing regional spatial and temporal variations in drought. Extensive time-series remote-sensing data were utilized, and we integrated the Temperature–Vegetation Dryness Index (TVDI), Drought Severity Index (DSI), Evaporation Stress Index (ESI), and the Temperature–Vegetation–Precipitation Dryness Index (TVPDI) to develop a comprehensive methodology for extracting regional vegetation drought characteristics. To mitigate the effects of regional drought non-stationarity on predictive accuracy, we propose a coupling-enhancement strategy that combines the Whale Optimization Algorithm (WOA) with the Informer model, enabling more precise forecasting of long-term regional drought variations. Unlike conventional deep-learning models, this approach introduces rapid convergence and global search capabilities, utilizing a sparse self-attention mechanism that improves performance while reducing model complexity. The results demonstrate that: (1) compared to the traditional Transformer model, test accuracy is improved by 43%; (2) the WOA–Informer model efficiently handles multi-objective forecasting for extended time series, achieving MAE (Mean Absolute Error) ≤ 0.05, MSE (Mean Squared Error) ≤ 0.001, MSPE (Mean Squared Percentage Error) ≤ 0.01, and MAPE (Mean Absolute Percentage Error) ≤ 5%. This research provides advanced predictive tools and precise model support for long-term vegetation restoration efforts.

Pixel-level Terrain Processing of the Martian Surface from Moderate Resolution Images of Tianwen-1

The moderate-resolution camera (MoRIC) of the Tianwen-1 orbiter, China’s first Mars exploration mission, is a frame-imaging scheme color camera. However, generating a precise digital elevation model (DEM) from these images faces challenges due to the presence of featureless and repetitive regions on the Martian surface. Traditional photogrammetric methods often struggle to achieve pixel-level resolution in these areas. To address this, we employ a shape from shading (SFS) algorithm that leverages shading patterns to reconstruct fine-scale terrain details, optimizing photogrammetric initial DEM. This paper presents a novel global bisection search algorithm, based on the Pearson correlation coefficient to determine the smoothing parameter, a crucial element in the SFS process. Comparisons with higher-resolution data from the Mars Express high-resolution stereo camera and Mars Reconnaissance Orbiter CTX images validate our approach, highlighting the potential of SFS to significantly enhance the scientific value of Tianwen-1 MoRIC data by generating high-resolution, three-dimensional maps of Mars.

Igniting the Flames of Entrepreneurial Intentions: Unravelling the Dynamics of Perceived Business Feasibility, Self-perceived Altruism, Innovativeness and Uncertainty Among University Students

The study seeks to elucidate the enablers shaping students’ entrepreneurial intentions across India’s top five central universities. The study evaluated the entrepreneurial intentions of 730 university students utilizing structural equation modelling and augmenting the theory of planned behaviour (TPB), the entrepreneurial event model (EEM) and the social entrepreneurship theory (SET). Mediation and moderation techniques facilitated the examination of the direct and indirect linkages in the model. The findings reveal that perceived behavioural control (PBC), entrepreneurial attitude, subjective norms and self-perceived altruism magnify perceived business feasibility (PBF), significantly escalating students’ entrepreneurial intentions. PBC and PBF partially mediate the relationship among the underlying constructs. In addition, perceived innovativeness reinforces the association, while perceived uncertainty weakens the association between PBF and students’ entrepreneurial intentions. The study’s findings could help policymakers, social entrepreneurs, educational institutions, impact investors and global development agencies strengthen India’s entrepreneurial ecosystem. Self-perceived altruism and PBF have not received adequate attention in entrepreneurship. We attempt to embrace these constructs and crucial personality traits by extending the horizons of TPB, EEM and SET using a precise conceptual model.

New generalized expressions for forced convective heat transfer coefficients across the flat plate

This study presents a novel generalized model for forced convective heat transfer coefficients (CHTC) over flat plates, addressing key limitations in existing models. Current expressions often provide averaged values and fail to capture the local variations in heat transfer, leading to inaccuracies, particularly in transitional zones. Moreover, there is no consensus on boundary layer classifications, which further complicates the development of reliable CHTC models. To address these issues, we conduct extensive CFD simulations to systematically analyze the combined effects of plate length (L), wind velocity (v), position along the plate (x), and temperature difference (Δt) on local CHTC. The simulations span a wide range of conditions (0.25 m ≤ L ≤ 8 m, 3 m/s ≤ v ≤ 25 m/s, and 5 °C ≤ Δt ≤ 80 °C), creating a comprehensive database of local CHTC values. Based on this data, a new generalized CHTC expression is derived as a function of L, v, x, and Δt, accurately predicting CHTC across laminar, transitional, and turbulent zones of flat plats. The accuracy of proposed CHTC expression is verified through thorough in-sample and out-of-sample evaluations, achieving R2 values between 0.973 and 0.999, with maximum deviations ranging from 2.6 % to 8.5 %. These research methods and conclusions can provide valuable references for optimizing thermal processes in applications such as industrial heat exchangers, electronics cooling, and building heating systems, where precise heat transfer modeling is critical for enhancing efficiency and system design.

Privacy enhancing and generalizable deep learning with synthetic data for mediastinal neoplasm diagnosis

The success of deep learning (DL) relies heavily on training data from which DL models encapsulate information. Consequently, the development and deployment of DL models expose data to potential privacy breaches, which are particularly critical in data-sensitive contexts like medicine. We propose a new technique named DiffGuard that generates realistic and diverse synthetic medical images with annotations, even indistinguishable for experts, to replace real data for DL model training, which cuts off their direct connection and enhances privacy safety. We demonstrate that DiffGuard enhances privacy safety with much less data leakage and better resistance against privacy attacks on data and model. It also improves the accuracy and generalizability of DL models for segmentation and classification of mediastinal neoplasms in multi-center evaluation. We expect that our solution would enlighten the road to privacy-preserving DL for precision medicine, promote data and model sharing, and inspire more innovation on artificial-intelligence-generated-content technologies for medicine.

Spatial-temporal Data Model for Indoor Fire Response Considering Multi-Factors

Abstract. The complexity of high-rise building structures and functions significantly increases the challenges in emergency responses to indoor fire incidents. This complexity is due to multiple factors that must be evaluated for decision-making within the fire scenario, such as the entities within the building, the status of environment, and the actions of people. An effective and precise indoor fire emergency management model should fully take into consideration of multiple factors, aiming to ensure the safety and effectiveness of emergency response action.This paper proposes a spatial-temporal data model to integrate the entity-status-action, for designing the most suitable action strategies in various environmental status by considering all related entities in the indoor fire scenario. Firstly, a list of factors is designed to represent the scenario features including spatial location, semantic description, attribute characteristics, geometric form, evolutionary process, and the relationships among objects. Then, employs an ontology-based descriptive method to integrate multi-dimensions and effectively represents the interrelations among entities, actions, and environmental status within the context of an indoor fire scenario. These environmental status considered in the model focus on the supporting path-planning during execution of direct response actions. Finally, a set of path-planning rules has been developed to infer a series of environmental status supporting to formulate the optimal evacuation strategy to minimize the risk of evacuation failure. The case study shows that the proposed model is effective in unified expression for multitude of factors associated with emergency response in indoor fire scenario and supporting the decision-making of the emergency response actions.

Deep learning-based anomaly detection using one-dimensional convolutional neural networks (1D CNN) in machine centers (MCT) and computer numerical control (CNC) machines

Computer numerical control (CNC) and machine center (MCT) machines are mechanical devices that manipulate different tools using computer programming as inputs. Predicting failures in CNC and MCT machines before their actual failure time is crucial to reduce maintenance costs and increase productivity. This study is centered around a novel deep learning-based model using a 1D convolutional neural network (CNN) for early fault detection in MCT machines. We collected sensor-based data from CNC/MCT machines and applied various preprocessing techniques to prepare the dataset. Our experimental results demonstrate that the 1D-CNN model achieves a higher accuracy of 91.57% compared to traditional machine learning classifiers and other deep learning models, including Random Forest (RF) at 89.71%, multi-layer perceptron (MLP) at 87.45%, XGBoost at 89.67%, logistic regression (LR) at 75.93%, support vector machine (SVM) at 75.96%, K-nearest neighbors (KNN) at 82.93%, decision tree at 88.36%, naïve Bayes at 68.31%, long short-term memory (LSTM) at 90.80%, and a hybrid 1D CNN + LSTM model at 88.51%. Moreover, our proposed 1D CNN model outperformed all other mentioned models in precision, recall, and F-1 scores, with 91.87%, 91.57%, and 91.63%, respectively. These findings highlight the efficacy of the 1D CNN model in providing optimal performance with an MCT machine’s dataset, making it particularly suitable for small manufacturing companies seeking to automate early fault detection and classification in CNC and MCT machines. This approach enhances productivity and aids in proactive maintenance and safety measures, demonstrating its potential to revolutionize the manufacturing industry.

Data-driven machine learning approaches for precise lithofacies identification in complex geological environments

ABSTRACT Reservoir characterization is a vital task within the oil and gas industry, with the identification of lithofacies in subsurface formations being a fundamental aspect of this process. However, lithofacies identification in complex geological environments with high dimensions, such as the Lower Indus Basin in Pakistan, poses a notable challenge, especially when dealing with limited data. To address this issue, we propose four common data-driven machine learning approaches: multi-resolution graph-based clustering (MRGC), artificial neural networks (ANN), K-nearest neighbors (KNN), and self-organizing map (SOM). We utilized these proposed approaches to assess their performance in scenarios with varying core sample availability, specifically evaluating their effectiveness in identifying lithofacies within the Lower Goru formation of the middle Indus Basin. The study reveals that in scenarios with a limited number of core samples, MRGC is the preferred choice, while KNN or MRGC is more suitable for larger datasets. The results demonstrate the superior performance of MRGC and KNN in lithofacies identification within the specified geological environment, with SOM following closely behind, and ANN exhibiting comparatively lower efficacy. The accurate identification of lithofacies from the selected model is complemented by the application of the truncated Gaussian simulation method for facies modeling. Comparative results confirm the excellent agreement between the model identification of lithofacies from well logs and electro-facies obtained from the truncated Gaussian simulation electro-facies volume. This study highlights the crucial role of selecting the right machine learning approach for precise lithofacies identification and modeling in complex geological environments. The comparative analysis provides practitioners in the petroleum industry with insights into the strengths and limitations of each method, enhancing existing knowledge. In conclusion, this research emphasizes the significance of comprehensive research and method selection for advancing lithofacies identification in diverse formations or study areas, ultimately benefiting the broader field of subsurface characterization in the petroleum industry.

Transported Entropy of Ions and Peltier Coefficients in 8YSZ and 10Sc1CeSZ Electrolytes for Solid Oxide Cells.

In this study, the transported entropy of ions for 8YSZ and 10Sc1CeSZ electrolytes was experimentally determined to enable precise modeling of heat transport in solid oxide cells (SOCs). The Peltier coefficient, crucial for thermal management, was directly calculated, highlighting reversible heat transport effects in the cell. While data for 8YSZ are available in the literature, providing a basis for comparison, the results for 10Sc1CeSZ show slightly smaller Seebeck coefficients but higher transported ion entropies. Specifically, at 700°C and an oxygen partial pressure of pO2=0.21 bar, values of SO2-*=52±10 J/K·F for 10Sc1CeSZ and SO2-*=48±9 J/K·F for 8YSZ were obtained. The transported entropy was also validated through theoretical calculations and showed minimal deviations when comparing different cell operation modes (O2||O2-||O2 and H2, H2O||O2-||O2). The influence of the transported entropy of the ions on the total heat generation and the partial heat generation at the electrodes is shown. The temperature has the greatest influence on heat generation, whereby the ion entropy also plays a role. Finally, the Peltier coefficients of 8YSZ for all homogeneous phases agree with the literature values.

Machine learning algorithms to predict treatment success for patients with pulmonary tuberculosis.

Despite advancements in detection and treatment, tuberculosis (TB), an infectious illness caused by the Mycobacterium TB bacteria, continues to pose a serious threat to world health. The TB diagnosis phase includes a patient's medical history, physical examination, chest X-rays, and laboratory procedures, such as molecular testing and sputum culture. In artificial intelligence (AI), machine learning (ML) is an advanced study of statistical algorithms that can learn from historical data and generalize the results to unseen data. There are not many studies done on the ML algorithm that enables the prediction of treatment success for patients with pulmonary TB (PTB). The objective of this study is to identify an effective and predictive ML algorithm to evaluate the detection of treatment success in PTB patients and to compare the predictive performance of the ML models. In this retrospective study, a total of 1236 PTB patients who were given treatment under a randomized controlled clinical trial at the ICMR-National Institute for Research in Tuberculosis, Chennai, India were considered for data analysis. The multiple ML models were developed and tested to identify the best algorithm to predict the sputum culture conversion of TB patients during the treatment period. In this study, decision tree (DT), random forest (RF), support vector machine (SVM) and naïve bayes (NB) models were validated with high performance by achieving an area under the curve (AUC) of receiver operating characteristic (ROC) greater than 80%. The salient finding of the study is that the DT model was produced as a better algorithm with the highest accuracy (92.72%), an AUC (0.909), precision (95.90%), recall (95.60%) and F1-score (95.75%) among the ML models. This methodology may be used to study the precise ML model classification for predicting the treatment success of TB patients during the treatment period.

AI in plastic surgery: customizing care for each patient

Artificial intelligence (AI) and machine learning (ML) involve the usage of complex algorithms to identify patterns, predict future outcomes, generate new data, and perform other tasks that typically require human intelligence. AI tools have been progressively adopted by multiple disciplines of surgery, enabling increasingly patient-specific care, as well as more precise surgical modeling and assessment. For instance, AI tools such as ChatGPT have been applied to enhance both patient educational materials and patient-surgeon communication. Additionally, AI tools have helped support pre- and postoperative assessment in a diverse set of procedures, including breast reconstructions, facial surgeries, hand surgeries, wound healing operations, and burn surgeries. Further, ML-supported 3D modeling has now been utilized for patient-specific surgical planning and may also be combined with 3D printing technologies to generate patient-customized, implantable constructs. Ultimately, the advent of AI and its intersection with surgical practice have demonstrated immense potential to transform patient care by making multiple facets of the surgical process more efficient, precise, and patient-specific.

Application of Fractional Modified Taylor Wavelets in the Dynamical Analysis of Fractional Electrical Circuits under Generalized Caputo Fractional Derivative

Abstract This study examines the application of fractional calculus in the analysis and modeling of electrical circuits of fractional order, highlighting its potential to explain the behaviour of complex electrical circuits accurately. In the domain of electrical circuits, fractional differential equations are employed in the analysis and simulation of systems that consist of resistors, capacitors and inductors. In the present paper, a novel approach utilizing fractional order modified Taylor wavelets is implemented to solve the fractional model of RL, LC, RC and RLC electrical circuits under generalized Caputo fractional derivative which offers precise and flexible modeling of non-locality and hereditary characteristics in complex systems. Furthermore, an additional parameter $\sigma$ (time scale parameter) is incorporated in fractional circuit dynamics to maintain the physical dimensionality. The considered wavelets with the collocation technique offer an efficient solution by converting the fractional model of electrical circuits into a set of algebraic equations which are further solved by using the Newton iteration method. Moreover, this study discusses the significance of Ulam-Hyers stability, emphasizing its role in ensuring stable and reliable circuit performance. The impact of fractional order on the dynamics of the electric circuit model is presented by tables and graphs. The approximate solutions obtained by the proposed method are well comparable with exact solutions and some other existing wavelet-based techniques. The residual errors are also evaluated under various model parameters for fractional orders. Furthermore, the graphs illustrate that the error progressively decreases as the number of wavelets basis increases.

Several suggestions for improving diagnosis and management of patients with neurofibromatosis type 1

Neurofibromatosis type 1 (NF1) is an autosomal dominant genetic disease caused by the mutations in the NF1 gene, with an incidence of approximately 1/3 000. Affecting multiple organs and systems throughout the body, NF1 caused a wide variety of clinical symptoms. A comprehensive multidisciplinary diagnostic and treatment model is needed to meet the diverse needs of NF1 patients and improve their quality of life. In recent years, the emergence of targeted therapies has further benefited NF1 patients, and the number of clinical consultations has increased dramatically. However, due to the rarity of the disease itself and insufficient attention previously, the standardized, systematic, and precise diagnosis and treatment model of NF1 still needs to be further improved. In this paper, we reviewed the current status of comprehensive diagnosis and treatment of NF1 in China, combine with our long-term experiences in diagnosis and treatment of this disease. Meanwhile, we propose future directions and several suggestions for the comprehensive diagnosis and treatment model for Chinese NF1 patients.

An open-access lumbosacral spine MRI dataset with enhanced spinal nerve root structure resolution

Spinal cord injury (SCI) profoundly affects an individual’s ability to move. Fortunately, recent advancements in neuromodulation, particularly the spatio-temporal epidural electrical stimulation (EES) targeting the spinal nerve roots, promoted rapid rehabilitation of SCI patients. Such neuromodulation techniques require precise anatomical modelling of spinal cord. However, the lack of spine imaging datasets, especially high-quality magnetic resonance imaging (MRI) datasets highlighting nerve roots, hinders the translation of EES into medical practice. To address this problem, we introduce an open-access lumbosacral spine MRI dataset acquired in 14 healthy adults, using constructive interference in steady state (CISS) sequence, double echo steady state (DESS) sequence, and T2-weight turbo spin echo (T2-TSE) sequence, with enhanced nerve root resolution. The dataset also includes the corresponding anatomical annotations of nerve roots and the final reconstructed 3D spinal cord models. The quality of our dataset is assessed using image quality metrics implemented in MRI quality control tool (MRIQC). Our dataset provides a valuable platform to promote a wide range of spinal cord neuromodulation research and collaboration among neurorehabilitation engineers.

Global mathematical analysis of a patchy epidemic model

The dissemination of a disease within a homogeneous population can typically be modeled and managed in a uniform fashion. Conversely, in non-homogeneous populations, it is essential to account for variations among subpopulations to achieve more precise predictive modeling and efficacious intervention strategies. In this study, we introduce and examine the comprehensive behavior of a deterministic two-patch epidemic model alongside its stochastic counterpart to assess disease dynamics between two heterogeneous populations inhabiting distinct regions. First, utilizing a specific Lyapunov function, we demonstrate that the disease-free equilibrium of the deterministic model is globally asymptotically stable. For the stochastic model, we establish that it is well-posed, meaning it possesses a unique positive solution with probability one. Subsequently, we ascertain the conditions necessary to ensure the total extinction of the disease across both regions. Furthermore, we explicitly determine a threshold condition under which the disease persists in both areas. Additionally, we discuss a scenario wherein the disease persists in one region while simultaneously becoming extinct in the other. The article concludes with a series of numerical simulations that corroborate the theoretical findings.

Adaptive control of an amplified piezoelectric-driven micropositioning stage based on notch filter structure inspiration and neural network online tuning

Abstract Piezoelectric-driven flexure micro/nanopositioning stages often exhibit a low-damping resonance mode, which can easily excite mechanical resonance during high-speed movement, and significantly impact the control system's stability, control bandwidth, and trajectory tracking accuracy. To mitigate the reliance on the precise modeling of stage dynamics inherent in current resonant controllers, an adaptive control method based on a back propagation (BP) neural network was designed to suppress resonance in real-time. First, a piezoelectric-driven flexure micro/nanopositioning stage system was constructed. Next, a feedback controller similar to a notch filter was designed, with bilinear transformation applied based on the system's inherent parameters to determine the initial values. Finally, the designed adaptive control method was tested through trajectory tracking experiments using a triangular wave signal. The experimental results showed that, when tracking the triangular wave signal, the maximum tracking error was reduced by 72.66% compared to proportional-integral (PI) control alone and by 68.66% compared to proportional integral control combined with a traditional notch filter. The tracking results demonstrate a significant improvement in the stage's stability and tracking accuracy.

Impact of Wettability on CO2 Dynamic Dissolution in Three-Dimensional Porous Media: Pore-Scale Simulation Using the Lattice Boltzmann Method.

Understanding the dissolution behavior of supercritical CO2 (scCO2) in porous media is crucial for efficient CO2 storage. However, the precise modeling of dynamic dissolution behavior at this pore scale remains a huge challenge, and the impact of wettability on this process still needs to be clarified. In this study, the influence of rock wettability on CO2 dynamic dissolution in the three-dimensional porous media is investigated using the lattice Boltzmann method (LBM). The LBM is coupled with scCO2-water two-phase flow, solute transport, and heterogeneous and homogeneous reactions. The size, number, and dissolution pattern of scCO2 bubbles during the dissolution process are observed under strongly water-wet, weakly water-wet, intermediate-wet, and mixed-wet conditions. The CO2(aq) concentration and pH are investigated, followed by a quantitative investigation of the impact of wettability on the specific interface area and the mass transfer coefficient. An empirical relationship between the specific interface area and scCO2 saturation is established. The findings reveal that under weakly water-wet and intermediate-wet conditions, the sizes of scCO2 clusters and monomers are small and mostly distributed at the dead end of the pores. In contrast, under strongly water-wet and mixed-wet conditions, the clusters are larger and interconnected, and distributed in the center of the pore. This results in a greater scCO2-water interface area, consequently enhancing the dissolution rate. Furthermore, a strong linear correlation is observed between scCO2 saturation and specific interface area. It is noted that as the hydrophilicity of the rock increases, the mass transfer coefficient initially rises and then declines.

Lightweight Detection of Broccoli Heads in Complex Field Environments Based on LBDC-YOLO

Robotically selective broccoli harvesting requires precise lightweight detection models to efficiently detect broccoli heads. Therefore, this study introduces a lightweight and high-precision detection model named LBDC-YOLO (Lightweight Broccoli Detection in Complex Environment—You Look Only Once), based on the improved YOLOv8 (You Look Only Once, Version 8). The model incorporates the Slim-neck design paradigm based on GSConv to reduce computational complexity. Furthermore, Triplet Attention is integrated into the backbone network to capture cross-dimensional interactions between spatial and channel dimensions, enhancing the model’s feature extraction capability under multiple interfering factors. The original neck network structure is replaced with a BiFPN (Bidirectional Feature Pyramid Network), optimizing the cross-layer connection structure, and employing weighted fusion methods for better integration of multi-scale features. The model undergoes training and testing on a dataset constructed in real field conditions, featuring broccoli images under various influencing factors. Experimental results demonstrate that LBDC-YOLO achieves an average detection accuracy of 94.44% for broccoli. Compared to the original YOLOv8n, LBDC-YOLO achieves a 32.1% reduction in computational complexity, a 47.8% decrease in parameters, a 44.4% reduction in model size, and a 0.47 percentage point accuracy improvement. When compared to models such as YOLOv5n, YOLOv5s, and YOLOv7-tiny, LBDC-YOLO exhibits higher detection accuracy and lower computational complexity, presenting clear advantages for broccoli detection tasks in complex field environments. The results of this study provide an accurate and lightweight method for the detection of broccoli heads in complex field environments. This work aims to inspire further research in precision agriculture and to advance knowledge in model-assisted agricultural practices.

Organoids-On-a-Chip for Personalized Precision Medicine.

The development of personalized precision medicine has become a pivotal focus in modern healthcare. Organoids-on-a-Chip (OoCs), a groundbreaking fusion of organoid culture and microfluidic chip technology, has emerged as a promising approach to advancing patient-specific treatment strategies. In this review, the diverse applications of OoCs are explored, particularly their pivotal role in personalized precision medicine, and their potential as a cutting-edge technology is highlighted. By utilizing patient-derived organoids, OoCs offer a pathway to optimize treatments, create precise disease models, investigate disease mechanisms, conduct drug screenings, and individualize therapeutic strategies. The emphasis is on the significance of this technological fusion in revolutionizing healthcare and improving patient outcomes. Furthermore, the transformative potential of personalized precision medicine, future prospects, and ongoing advancements in the field, with a focus on genomic medicine, multi-omics integration, and ethical frameworks are discussed. The convergence of these innovations can empower patients, redefine treatment approaches, and shape the future of healthcare.

Evaluation of Selected Factors Affecting the Speed of Drivers at Signal-Controlled Intersections in Poland

In traffic engineering, vehicle speed is a critical determinant of both the risk and severity of road crashes, a fact that holds particularly important for signalized intersections. Accurately selecting vehicle speeds is crucial not only for minimizing accident risks but also for ensuring the proper calculation of intergreen times, which directly influences the efficiency and safety of traffic flow. Traditionally, the design of signal programs relies on fixed speed parameters, such as the posted speed limit or the operational speed, typically represented by the 85th percentile speed from speed distribution data. Furthermore, many design guidelines allow for the selection of these critical speed values based on the designer’s own experience. However, such practices may lead to discrepancies in intergreen time calculations, potentially compromising safety and efficiency at intersections. Our research underscores the substantial variability in the speeds of passenger vehicles traveling intersections under free-flow conditions. This study encompassed numerous intersections with the highest number of accidents, using unmanned aerial vehicles to conduct surveys in three Polish cities: Toruń, Bydgoszcz, and Warsaw. The captured video footage of vehicle movements at predetermined measurement sections was analyzed to find appropriate speeds for various travel maneuvers through these sections, encompassing straight-through, left-turn, and right-turn relations. Our analysis focused on how specific infrastructure-related factors influence driver behavior. The following were evaluated: intersection type, traffic organization, approach lane width, number of lanes, longitudinal road gradient, trams or pedestrian or bicycle crossing presence, and even roadside obstacles such as buildings, barriers or trees, and others. The results reveal that these factors significantly affect drivers’ speed choices, particularly in turning maneuvers. Furthermore, it was observed that the average speeds chosen by drivers at signalized intersections did not reach the permissible speed limit of 50 km/h as established in typical Polish urban areas. A key outcome of our analysis is the recommendation for a more precise speed model that contributes to the design of signal programs, enhancing road safety, and aligning with sustainable transport development policies. Based on our statistical analyses, we propose adopting a more sophisticated model to determine actual vehicle speeds more accurately. It was proved that, using the developed model, the results of calculating the intergreen times are statistically significantly higher. This recommendation is particularly pertinent to the design of signal programs. Furthermore, by improving speed accuracy values in intergreen calculation models with a clear impact on increasing road safety, we anticipate reductions in operational costs for the transportation system, which will contribute to both economic and environmental goals.