Maximum Intersection Research Articles

Improved semantic segmentation method for weld penetration prediction of TIG welding with dual ellipsoid heat source

TIG welding is a manufacturing process that uses argon as a shielding gas and tungsten as an electrode to join metals at high temperatures. The weld penetration is the key to determining the quality of the weld. However, the lack of sensing technology makes weld penetration difficult to predict, which imposes a major challenge to process stability and weld quality. To address this challenge, a semantic segmentation-based approach for calibrating the weld penetration prediction model under dual ellipsoidal heat source is proposed to improve the prediction accuracy. To realize this, a semantic segmentation model Self-Attention-Structural-Reparameterization-BiSeNet (SASR-BiSeNet) is built for molten pool parameter extraction. The model introduces a self-attention mechanism as an enhancement to the convolution module. Meanwhile, the RepVGG-style structural reparameterization design decouples the training-time and inference-time architectures. Further, a mask feature extraction method is designed to obtain the calibration parameters of the molten pool and to correct the prediction model. The accuracy of the calibrated prediction model is verified by welding experiments using 304L steel materials on a robotic welding system. The results show that the maximum Mean Intersection over Union (mIoU) of the SASR-BiSeNet is 92.51 %, which is a 13.87 % improvement compared to the BiSeNet, and the maximum depth of molten pool error decreasing from 16.59 % to 9.84 %. This method can improve the precision of welding penetration control and plays an important role in promoting welding intelligence.

Description of the spatial variability of concrete via composite random field and failure analysis of chimney

The inherent variability of concrete significantly affects the structural safety and performance. The variability of concrete is a complex phenomenon influenced by multiple factors, including material properties, production processes, and environmental conditions. Understanding and quantifying the variability of concrete is crucial for reliable and safe structural design. Probabilistic methods are commonly used to account for concrete variability in structural design. In this paper, a composite random field approach combined with a hierarchy model is used to consider the multi-scale spatial variability of concrete. The random field of compressive strength is expressed as a sum of independent component random fields. To investigate the impact of concrete's spatial variability on structural response and failure modes, the failure analysis of a 115-m-tall chimney was conducted. The results indicate that the composite random field approach proves to be a valuable method for incorporating concrete's spatial variability at different scales. The spatial variability of concrete exerts a substantial influence on the potential positions where severe compressive damage might occur. Additionally, the failure modes are also affected by the spatial variability of concrete. When taking into account the spatial variability of concrete, an extra collapse mode emerges, aligning more closely with the chimney's actual collapse mode during an earthquake. Furthermore, the spatial variability of concrete also moderately impacts the variability of the base shear force and the maximum inter-section drift angle. Notably, improper approaches to considering the spatial variability of concrete significantly impact the concrete's compressive damage and structural response.

Introducing vacancy defects by rapid quenching in FCC metal for deep learning micrograph dataset

Accurate detection of crystal defects from microstructure micrographs is essential for evaluating the mechanical and functional properties of metallic materials. Emerging deep learning techniques have great potential for application in crystal defects detection, but the excellent detection accuracy requires domain knowledge and micrograph datasets. Here we target to detect and distinguish nanoscale dislocation loops (DLs) and stacking fault tetrahedrons (SFTs) in transmission electronic microscope (TEM) micrographs. The micrograph datasets of crystal defects derive from copper prepared by rapid quenching and post-annealing. Applying standard U-Net architecture, the maximum Mean Intersection over Union (mIOUmax) of 3-class model is 54.29%, while it is 73.81% and 74.67% of DL 2-class and SFT 2-class model. Results suggest that excellent detection accuracy of deep learning model could be achieved by learning features from specific defect datasets. These findings provide a new strategy for accurate detection of crystal defects in the field of material engineering.

Fast registration method for sequential star images.

Registration of sequential star images is an important component of space observation. Especially for onboard devices, the efficiency and robustness of registration algorithms are particularly important. The typical star image registration approach based on star matching has a low tolerance for incorrect matching, and the time cost will increase rapidly as the number of stars increases. Due to the high overlap and rigid transformation of adjacent star images, a proposed method based on star angular distance (SAD) is presented. Stars are easy to locate and extract as natural feature points, and there are a large number of identical stars in adjacent star images. The rotation and translation of the SAD, composed of identical stars in adjacent star images, are the same. Therefore, maximum intersection clustering (MIC) was proposed to cluster rotation and translation, and Gaussian weight iteration (GWI) was proposed to estimate rigid transformation parameters. The use of SAD as a star image feature reduces the complexity of star image features, which can improve the efficiency of the algorithm. MIC can tolerate errors within a certain range, and GWI can lessen their impact on the results, increasing the algorithm's robustness. Experimental results show that the proposed method can improve the trend of rapidly increasing computation as the number of stars increases and avoid the restriction that transformation parameters must be obtained with correctly matching stars. Compared to the typical triangle method and SAD similarity method, the proposed method has higher efficiency under different numbers of stars, and translation, rotation, and location errors.

Sequence Reconstruction Under Single-Burst-Insertion/Deletion/Edit Channel

Motivated by applications in modern storage devices such as DNA storage and racetrack memories, we study the sequence reconstruction problem which involves two important issues. One is determining the maximum intersection size between two error balls. The other is designing reconstruction codes in which each transmitted sequence can be reconstructed from a given number of noisy reads with redundancy as small as possible. In this paper, we focus on channels that introduce single-burst-insertion/ deletion error. We fully determine the maximum intersection size between two burst-insertion/deletion balls for both fixed-length and variable-length models. Then we characterize a pair of sequences when their burst-insertion/deletion balls have intersection of a certain size. Using these characterizations, we design reconstruction codes for all values <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i> of the number of noisy reads and analyze their lower bounds. In addition, we extend our results to the burst-edit channel.

Analisis Struktur Baja dengan Perhitungan Beban Gempa Menggunakan Metode Time History Berdasarkan SNI 1726:2012 dan SNI 1729:2015

The selection of material types used in designing a building is one of essential parts since every type of material consisted of different elements. One of primary materials in constructing building structure is steel. The use of steel has been implemented in most building structures nowadays for its high performance of ductility compared to concrete and wood. Ductility is an important requirement to achieve building structure that is resistant to earthquake. In this thesis, there will be an elaboration of an earthquake resistant and steel structured building. The designed building will be a five-floor office configured by an organized structure. The steel framework system of specialized moment loader is used as structure system of earthquake resistance and analysed using time history method. The steel material used has A36 grade (fy = 250 MPa; fu = 400 MPa), and the concrete grade f’c = 30 MPa. The preparation process results in a design of an earthquake steel resistant structure building that meets the requirement of intersection between each floor and the maximum intersection value occurs on second floor with 93,5 mm where the legal supposed intersection has been valued on 95 mm. The dimension of structure element used is profile block W 14 x 6,75 x 38 for the x-axis and W 16 x 7 x 40 for y-axis. It also has 12 cm of plate thickness on every level with profile column of Wx14x16x283

ULMFiT Embedding(s) for Context and Extended Gloss Intersection for Marathi Word Sense Disambiguation

Ambiguities in the word meanings makes all the natural language processing (NLP) tasks very difficult, word sense disambiguation (WSD) is used to resolve these ambiguities. Now a day’s NLP-based human assistive systems are in demand, in which machines are expected to resolve word sense ambiguities. Today, due to the availability of machine readable dictionaries knowledge-based WSD approaches have become popular; it explores semantic relations between the contextual features and possible glosses of the given ambiguous word. Inductive transfer learning-based language models have great potential to represent the different semantic features of the word, which can be used in various NLP tasks. Universal language model fine-tuning for text classification (ULMFiT) is a popular transfer learning model used to embed various semantic features in digitally resource scare and morphologically rich language like marathi. In this reported work, the ambiguous words from the Marathi input sentence is extracted and have obtained its possible synset and glosses from IndoWordNet, these glosses are then extended using hypernym and hyponym relations. We have obtained the word embedding of marathi context and extended glosses using ULMFiT model. For the test run, we have crafted the test-bed of 6000 marathi sentences of 280 moderately ambiguous words harvested from marathi websites, which caters for 1200 senses. The winner sense is declared based on the maximum intersection score between the pair of context and gloss embedding. We have obtained the average accuracy up to 57.10% for our dataset.

Obtaining the Dimensions and Orientation of 2D Rectangular Flakes from Sectioning Experiments in Flake Composites

Recently, we developed and reported the statistical validity of two methods for determining the planar aspect ratios of two-dimensional (2D) rectangular flakes in composites from the statistics of intersection lengths: one method is based on the maximum intersection length, and the other on the average intersection length. In this work, we show that these methods are valid and robust not only for flakes having isotropic, random in-plane orientations, but for the more general situations of planar orientations ranging from unidirectional (misalignment angle ϵ=0), to partially aligned (0&lt;ϵ&lt;π/2), to flakes of isotropic, random-in-plane orientations (ϵ=π/2). We prove, by Monte Carlo simulations and by numerical sectioning experiments, the validity of the proposed methods for characterizing the extent of the partial alignment (the misalignment angle ϵ) of 2D rectangular flakes in composites, based again on the statistics of the intersection lengths; this information can be obtained from cross-sections of composite samples used in optical or electron microscopy or using tomographic imaging techniques. The performance of these techniques was tested using blind experiments in numerically sectioned composites which contained up to 106 individual flakes, and was found to be very good for a wide range of flake aspect ratios.

Preliminary study for developing a navigation system for gastric cancer surgery using artificial intelligence.

We are attempting to develop a navigation system for safe and effective peripancreatic lymphadenectomy in gastric cancer surgery. As a preliminary study, we examined whether or not the peripancreatic dissection line could be learned by a machine learning model (MLM). Among the 41 patients with gastric cancer who underwent radical gastrectomy between April 2019 and January 2020, we selected 6 in whom the pancreatic contour was relatively easy to trace. The pancreatic contour was annotated by a trainer surgeon in 1242 images captured from the video recordings. The MLM was trained using the annotated images from five of the six patients. The pancreatic contour was then segmented by the trained MLM using images from the remaining patient. The same procedure was repeated for all six combinations. The median maximum intersection over union of each image was 0.708, which was higher than the threshold (0.5). However, the pancreatic contour was misidentified in parts where fatty tissue or thin vessels overlaid the pancreas in some cases. The contour of the pancreas could be traced relatively well using the trained MLM. Further investigations and training of the system are needed to develop a practical navigation system.

Reactive medullary astrocytes are associated with RVLM neuroinflammation in heart failure rats

Heart failure with preserved ejection fraction (HFpEF) is a complex clinical syndrome. Brain inflammation has been suggested to play a key role in HFpEF progression by modulating the activity of several brainstem circuits encompassing cardiovascular and respiratory regulation. Indeed, we recently reported RVLM‐C1 catecholaminergic neurons hyperactivity in rats with HFpEF. However, the precise mechanism underpinning changes in C1 activity in HFpEF remains to be determined. Interestingly, glial cell activation within the RVLM (particularly astrocytes) has been shown to modulate C1 neuron activity. Despite this evidence whether astrocytes may become activated in HFPEF has not been previously studied. Accordingly, we aimed to determine changes in astrocyte activity within the RVLM of HFpEF rats and its relation with RVLM neuroinflammatory status. Adult male Sprague‐Dawley rats (250 g) underwent volume overload to induce HFpEF. Eight weeks after surgical induction of HFpEF, rats underwent respiratory function recordings in freely moving conditions using whole‐body plethysmography. At the end of the physiological experiments, brains were harvested, fixed in 4% PFA and processed for immunohistochemical analysis. Immunofluorescence for tyrosine hydroxylase and glial fibrillary acidic protein (GFAP) were performed in 25 μm free‐floating brainstem sections containing the RVLM. Sholl analysis was performed to evaluate astrocytic morphological differences between HFpEF and Sham rats (total length of processes, # of processes, maximum intersection radius, # of intersections). RVLM neuroinflammatory status was measured in micropunches by immunoblot (GFAP) and qPCR (TNF‐α, IL‐1β). Compared to Sham rats, HFpEF animals showed clear signs of astrocyte reactivity evidenced by increases in the number of astrocyte processes (11.27 ± 1.096 vs. 18.64 ± 1.333; HFpEF vs. Sham, respectively), length of astrocyte processes (267.5 ± 17.92μm vs. 392.8 ± 22.53μm; HFpEF vs. Sham, respectively) and number of intersections (3.410 ± 0,25 vs. 4.538 ± 0.317; HFpEF vs. Sham, respectively). Consistently, we found an increased GFAP expression within the RVLM of HFpEF rats compared with Sham rats (99.9 ± 5.9 vs. 135.3 ± 15.8; HFpEF vs. Sham, respectively. In addition, we found that TNF‐α and IL‐1β were both augmented in HFpEF (~7 and ~3 fold, respectively) compared to Sham. Notably, respiratory dysfunction observed in HFpEF were significantly and positively correlated with the degree of activation of RVLM astrocytes. Indeed, respiratory irregularity score (R2 = 0.98; p&lt;0.05), short‐term variability of the breath‐to‐breath interval (R2 = 0.98; p&lt;0.05), and apneas/hypopneas index (R2 = 0.94; p&lt;0.05)) were all related to GFAP expression in HFpEF rats. Our findings showed the presence of reactive astrocytes within the RVLM of HFpEF animals and that the degree of astrocyte activity was associated with RVLM neuroinflammation. Together our results suggest that increases in astrocyte reactivity in HFpEF may help to induce/maintain a pro‐iflammatory niche within the RVLM that ultimately lead to C1 neuron activity alterations.

A Statistical Shape Model-Based Analysis of Periacetabular Osteotomies: Technical Considerations to Achieve the Targeted Correction.

Classic and reverse Bernese periacetabular osteotomy (PAO) have been shown to be effective for the treatment of developmental dysplasia of the hip (by classic PAO), severe acetabular retroversion (by reverse PAO), and some protrusio acetabuli (by reverse PAO). Especially in severe cases with higher degrees of correction, a relevant overlap between the osteotomized fragment and the pelvis might occur, leading to necessary fragment translation. The aim of the present study was to analyze the necessary translation as a function of the degree of correction using a statistical mean model of the pelvis according to the technique (classic PAO or reverse PAO). A mean statistical shape model of the pelvis and 2 extreme models were used to simulate rotation of the osteotomized fragment during a classic or reverse PAO and to calculate rotations from -20° to 20° in the frontal, sagittal, and transverse planes and a combination thereof. The depth and volume of the intersection between the mobilized fragment and the pelvis were calculated, and the minimum translation of the fragment necessary to avoid segment overlap was determined. The maximum intersection distances between the pelvis and the 20° rotated fragment were 6.7 and 15.3 mm for adduction and abduction (frontal plane), 6.4 and 4.5 mm for internal and external rotation (transverse plane), and 27.8 and 9.2 mm for extension and flexion (sagittal plane). The necessary translations for 20° of fragment rotation were 7.0 and 12.8 mm for adduction and abduction (frontal plane), 4.8 and 5.0 mm for internal and external rotation (transverse plane), and 18.5 mm and 8.8 mm for extension and flexion (sagittal plane). Acetabular reorientation with the classic or reverse PAO results in translation of the fragment and in a consequent change in the rotational center. This finding is more pronounced with higher degrees of fragment reorientation in abduction and extension; it becomes especially pronounced in reverse PAO for acetabular retroversion or protrusio acetabuli, and might limit the ability to achieve the intended improvement in overall hip biomechanics.

A Novel Method for the Determination of the Lateral Dimensions of 2D Rectangular Flakes.

We present a novel method for the determination of the lateral dimensions of thin rectangular flakes, as they exist randomly dispersed in flake composites. Knowledge of flake size and shape is essential for the correct prediction of the mechanical, electrical, thermal and barrier properties of flake composites. The required information is the distribution function of lengths of the lines representing the intersection of flakes with a sectioning plane, as seen in cross-sections of composite samples used in optical or electron microscopy or obtained using tomographic imaging techniques. The key observation is that the major peak of the distribution function coincides with the short dimension S of the flake while a secondary peak corresponds to its long dimension W. These observations are explained using Monte-Carlo simulations, as well as deterministic, geometry-based modeling and probability analysis. Since the strength of the secondary peak diminishes with increasing flake aspect ratio , we develop two additional methods for the determination of W. The first finds W from the maximum intersection length; this procedure is justified by computing the relevant probability fields through Monte-Carlo simulations. The second method finds r from the average intersection length and is valid in the range . The performance of these techniques is tested and found to be very good using blind experiments in numerically sectioned specimens.

A GRASP algorithm with Tabu Search improvement for solving the maximum intersection of k-subsets problem

The selection of individuals with similar characteristics from a given population have always been a matter of interest in several scientific areas: data privacy, genetics, art, among others. This work is focused on the maximum intersection of k-subsets problem (kMIS). This problem tries to find a subset of k individuals with the maximum number of features in common from a given population and a set of relevant features. The research presents a Greedy Randomized Adaptive Search Procedure (GRASP) where the local improvement is replaced by a complete Tabu Search metaheuristic with the aim of further improving the quality of the obtained solutions. Additionally, a novel representation of the solution is considered to reduce the computational effort. The experimental comparison carefully analyzes the contribution of each part of the algorithm to the final results as well as performs a thorough comparison with the state-of-the-art method. Results, supported by non-parametric statistical tests, confirms the superiority of the proposal.

Improving Road Surface Area Extraction via Semantic Segmentation with Conditional Generative Learning for Deep Inpainting Operations

The road surface area extraction task is generally carried out via semantic segmentation over remotely-sensed imagery. However, this supervised learning task is often costly as it requires remote sensing images labelled at the pixel level, and the results are not always satisfactory (presence of discontinuities, overlooked connection points, or isolated road segments). On the other hand, unsupervised learning does not require labelled data and can be employed for post-processing the geometries of geospatial objects extracted via semantic segmentation. In this work, we implement a conditional Generative Adversarial Network to reconstruct road geometries via deep inpainting procedures on a new dataset containing unlabelled road samples from challenging areas present in official cartographic support from Spain. The goal is to improve the initial road representations obtained with semantic segmentation models via generative learning. The performance of the model was evaluated on unseen data by conducting a metrical comparison where a maximum Intersection over Union (IoU) score improvement of 1.3% was observed when compared to the initial semantic segmentation result. Next, we evaluated the appropriateness of applying unsupervised generative learning using a qualitative perceptual validation to identify the strengths and weaknesses of the proposed method in very complex scenarios and gain a better intuition of the model’s behaviour when performing large-scale post-processing with generative learning and deep inpainting procedures and observed important improvements in the generated data.

TREASURE: Text Mining Algorithm Based on Affinity Analysis and Set Intersection to Find the Action of Tuberculosis Drugs against Other Pathogens

Tuberculosis (TB) is one of the top causes of death in the world. Though TB is known as the world’s most infectious killer, it can be treated with a combination of TB drugs. Some of these drugs can be active against other infective agents, in addition to TB. We propose a framework called TREASURE (Text mining algoRithm basEd on Affinity analysis and Set intersection to find the action of tUberculosis dRugs against other pathogEns), which particularly focuses on the extraction of various drug–pathogen relationships in eight different TB drugs, namely pyrazinamide, moxifloxacin, ethambutol, isoniazid, rifampicin, linezolid, streptomycin and amikacin. More than 1500 research papers from PubMed are collected for each drug. The data collected for this purpose are first preprocessed, and various relation records are generated for each drug using affinity analysis. These records are then filtered based on the maximum co-occurrence value and set intersection property to obtain the required inferences. The inferences produced by this framework can help the medical researchers in finding cures for other bacterial diseases. Additionally, the analysis presented in this model can be utilized by the medical experts in their disease and drug experiments.

Convolutional neural networks for water segmentation using sentinel-2 red, green, blue (RGB) composites and derived spectral indices

ABSTRACT Near-real time water segmentation with medium resolution satellite imagery plays a critical role in water management. Automated water segmentation of satellite imagery has traditionally been achieved using spectral indices. Spectral water segmentation is limited by environmental factors and requires human expertise to be applied effectively. In recent years, the use of convolutional neural networks (CNN’s) for water segmentation has been successful when used on high-resolution satellite imagery, but to a lesser extent for medium resolution imagery. Existing studies have been limited to geographically localized datasets and reported metrics have been benchmarked against a limited range of spectral indices. This study seeks to determine if a single CNN based on Red, Green, Blue (RGB) image classification can effectively segment water on a global scale and outperform traditional spectral methods. Additionally, this study evaluates the extent to which smaller datasets (of very complex pattern, e.g harbour megacities) can be used to improve globally applicable CNNs within a specific region. Multispectral imagery from the European Space Agency, Sentinel-2 satellite (10 m spatial resolution) was sourced. Test sites were selected in Florida, New York, and Shanghai to represent a globally diverse range of waterbody typologies. Region-specific spectral water segmentation algorithms were developed on each test site, to represent benchmarks of spectral index performance. DeepLabV3-ResNet101 was trained on 33,311 semantically labelled true-colour samples. The resulting model was retrained on three smaller subsets of the data, specific to New York, Shanghai and Florida. CNN predictions reached a maximum mean intersection over union result of 0.986 and F1-Score of 0.983. At the Shanghai test site, the CNN’s predictions outperformed the spectral benchmark, primarily due to the CNN’s ability to process contextual features at multiple scales. In all test cases, retraining the networks to localized subsets of the dataset improved the localized region’s segmentation predictions. The CNN’s presented are suitable for cloud-based deployment and could contribute to the wider use of satellite imagery for water management.

The robust minimal controllability and observability problem

AbstractIn this paper, we study the Robust Minimal Controllability and Observability Problem (rMCOP). The scenario that motivated this question is related to the design of a drone formation to execute some task, where the decision of which nodes to equip with a more expensive communication system represents a critical economic choice. Given a linear time‐invariant system for each of the vehicles, this problem consists of identifying a minimal subset of state variables to be actuated and measured, ensuring that the overall formation model is both controllable and observable while tolerating a prescribed level of inputs/outputs that can fail. Based on the tools in the available literature, a naive approach would consist of enumerating separately all possible minimal solutions for the controllability and observability parts. Then, iterating over all combinations to find the maximum intersection of sensors/actuators in the independent solutions, yielding a combinatorial problem. The presented solution couples the design of both controllability and observability parts through a polynomial reformulation as a minimum set multi‐covering problem under some mild assumptions. In this format, the algorithm has the following interesting attributes: (i) only requires the solution of a single covering problem; 9ii) using polynomial approximations algorithms, one can obtain close‐to‐optimal solutions to the rMCOP.

Emergence and synchronization of a reversible core in a system of forced adaptively coupled Kuramoto oscillators.

We report on the phenomenon of the emergence of mixed dynamics in a system of two adaptively coupled phase oscillators under the action of a harmonic external force. We show that in the case of mixed dynamics, oscillations in forward and reverse time become similar, especially at some specific frequencies of the external force. We demonstrate that the mixed dynamics prevents forced synchronization of a chaotic attractor. We also show that if an external force is applied to a reversible core formed in an autonomous case, the fractal dimension of the reversible core decreases. In addition, with increasing amplitude of the external force, the average distance between the chaotic attractor and the chaotic repeller on the global Poincaré secant decreases almost to zero. Therefore, at the maximum intersection, we see a trajectory belonging approximately to a reversible core in the numerical simulation.

Optimization design of O-ring linear accelerator gantry and mechanical isocenter detection method.

With the development of the precise radiotherapy, the accuracy of radiotherapy equipment is gradually improved. The gantry, the carrier of the treatment head and various testing devices, is the most important component that determines the accuracy of the entire equipment. In this paper, the layout of the O-ring linear accelerator and the structure of the gantry are optimized to reduce the weight of the gantry by 50% and the moment of inertia by 60%. A mechanical isocenter detection method based on the laser tracker is proposed to conduct a rapid and accurate isocenter measurement for the optimized gantry. The experimental results show that the maximum deformation of the optimized gantry under load is 0.13mm during rotation and the maximum intersection distance between the gantry axis and the treatment head axis at the isocenter position is 0.21mm.

УСЛОВИЕ БЕЗАВАРИЙНОГО ДВИЖЕНИЯ АВТОМОБИЛЕЙ НА РЕГУЛИРУЕМЫХ ПЕРЕСЕЧЕНИЯХ

Introduction: Improving intersection capacity will not be possible without accounting for traffic safety. Purpose of the study: We aim to determine the prerequisite for accident-free traffic at signal-controlled intersections with turning traffic flows. Methods: In our study, we used the methods of observation, comparison, and mathematical analysis. Results: We have carried out a field observation of traffic intensity at signal-controlled intersections in the city of Yekaterinburg, focusing on vehicles that moved when the green light was on. We have also analyzed traffic flow moving in three directions in the same lanes. We have discovered that traffic accident likelihood is the highest (54%) at four-way intersections. Three-way intersections account for 44% of traffic accidents, while the remaining 2% of accidents occur at multi-way intersections. Furthermore, we have determined the additional factors that impact safety in turning traffic flows at intersections. Our study demonstrates that in order to ensure maximum intersection capacity, the duration of the traffic signal cycle must be adjusted with the minimum safe distance between vehicles in mind.