Test Scenarios Research Articles

MGFNet: Cross-scene Crowd Counting via Multistage Gated Fusion Network

Existing crowd counting methods are mainly trained and tested in similar scenarios. When the testing and training scenarios of the model are different, the counting accuracy of these methods will sharply decrease, which seriously limits their practical application. To address this problem, we propose a multistage gated fusion network (MGFNet) for cross-scene crowd counting. MGFNet is primarily composed of dynamic gated convolution units (DGCU) and multilevel scale attention blocks (MSAB) modules. Specifically, DGCU uses a dynamic gating path to supplement detailed information to reduce the loss of crowd information and overestimation of background in different scenarios. MSAB calibrates crowd information at different scales and perspectives in different scenes by generating attention maps with discriminative information. In addition, we used a new global local consistency loss to optimize the model to adapt to changes in crowd density and distribution. Extensive experiments on four different types of scene counting benchmarks show that the proposed MGFNet achieves superior cross-scene counting performance.

Risk and Complexity Assessment of Autonomous Vehicle Testing Scenarios

Autonomous vehicles (AVs) must fulfill adequate safety requirements before formal application, and performing an effective functional evaluation to verify vehicle safety requires extensive testing in different scenarios. However, it is crucial to rationalize the application of different scenarios to support different testing needs; thus, one of the current challenges limiting the development of AVs is the critical evaluation of scenarios, i.e., the lack of quantitative criteria for scenario design. This study introduces a method using the Spherical Fuzzy-Analytical Network Process (SF-ANP) to evaluate these scenarios, addressing their inherent risks and complexities. The method involves constructing a five-layer model to decompose scenario elements and using SF-ANP to calculate weights based on element interactions. The study evaluates 700 scenarios from the China In-depth Traffic Safety Study–Traffic Accident (CIMSS-TA) database, incorporating fuzzy factors and element weights. Virtual simulation of vehicles in the scenarios was performed using Baidu Apollo, and the performance of the scenarios was assessed by collecting the vehicle test results. The correlation between the obtained alternative safety indicators and the quantitative values confirms the validity and scientific validity of this approach. This will provide valuable guidance for categorizing audiovisual test scenarios and selecting corresponding scenarios to challenge different levels of vehicle functionality. At the same time, it can be used as a design basis to generate a large number of effective scenarios to accelerate the construction of scenario libraries and promote commercialization of AVs.

Simulating Pelvis Kinematics from Belt and Seat Loading in Frontal Car Crash Scenarios: Important Boundary Conditions that Influence the Outcome.

The risk of submarining during automotive crashes, defined by the lap belt sliding off the pelvis to load the abdomen, is predicted to increase in future autonomous vehicles as greater variation in seating position is enabled. Biofidelic tools are required to efficiently design and evaluate new and/or improved safety systems. This study aims to evaluate the pelvis response sensitivity to variations in boundary conditions that directly influence the pelvis loads, deemed important for the submarining outcome, to facilitate a more precise comparison between finite element human body models (FE-HBMs) and post-mortem human subjects (PMHSs). A parameter study, using a one-variable-at-a-time analysis (low/high) of belt friction, seat friction, seat stiffness, and (on/off) for added belt bending stiffness, was performed using a state-of-the-art FE-HBM in four different test scenarios; one stationary, two sleds with upright occupant posture, and one sled with reclined occupant posture. In the stationary scenario, both belt friction and belt bending stiffness influenced the belt folding behavior, which consequently affected the belt-to-pelvis angle at submarining. In the sled scenarios, only seat friction was found to influence the pelvis kinematics and submarining outcome, with the most biofidelic response resulting from both the low (0.2) and high (0.5) friction coefficient depending on the scenario. To reduce uncertainty in boundary conditions affecting the external pelvis loads and increase confidence in FE-HBM to PMHS comparisons, it is recommended that future experiments evaluate the PMHS to seat friction coefficient and that new belt modeling methods that accurately capture belt folding when interacting with soft tissues are developed.

Behavioral Cloning Strategies in Steering Angle Prediction: Applications in Mobile Robotics and Autonomous Driving

Artificial neural networks (ANNs) are artificial intelligence techniques that have made autonomous driving more efficient and accurate; however, autonomous driving faces ongoing challenges in the accuracy of decision making based on the analysis of the vehicle environment. A critical task of ANNs is steering angle prediction, which is essential for safe and effective navigation of mobile robots and autonomous vehicles. In this study, to optimize steering angle prediction, NVIDIA’s architecture was adapted and modified along with the implementation of the Swish activation function to train convolutional neural networks (CNNs) by behavioral cloning. The CNN used human driving data obtained from the UDACITY beta simulator and tests in real scenarios, achieving a significant improvement in the loss function during training, indicating a higher efficiency in replicating human driving behavior. The proposed neural network was validated through implementation on a differential drive mobile robot prototype, by means of a comparative analysis of trajectories in autonomous and manual driving modes. This work not only advances the accuracy of steering angle predictions but also provides valuable information for future research and applications in mobile robotics and autonomous driving. The performance results of the model trained with the proposed CNN show improved accuracy in various operational contexts.

Evaluation and Revision of Historical ESSENCE Syndromic Surveillance Definition Used to Identify Work-Related Emergency Department Visits.

Occupational health surveillance programs have rarely incorporated syndromic surveillance. We documented a process for evaluating and revising syndromic surveillance definitions, using the example of work-related emergency department (ED) visits. Of 140 331 481 ED visits in the National Syndromic Surveillance Program from May 15, 2016, through October 31, 2017, we extracted 391 827 records for phase 1 that met the previously unevaluated original work-related definition or included a work-related employment status ICD-10-CM code (Y99.0/Y99.1; International Classification of Diseases, Tenth Revision, Clinical Modification). Also for phase 1, we evaluated the performance of the original work-related definition using ICD-10-CM work-related employment status codes as a reference standard. In phase 2, we extracted 99 109 records containing candidate work-related keywords. Finally, in phase 3, we examined individual keywords and developed and evaluated a revised definition. The revised work-related definition identified 259 153 visits by keyword alone, 104 342 visits by ICD-10-CM work-related employment status codes alone, and 16 385 ED visits by both methods. Sensitivity (13.6% vs 11.3%) and positive predictive value (5.9% vs 4.8%) were higher for the revised definition than for the original work-related definition. Manual review indicated that 91.3% (95% CI, 89.6%-93.0%) of visits identified by the revised definition keywords met the study criteria for work-related visits as compared with 75.5% (95% CI, 73.0%-78.4%) of visits for the original work-related definition. The original work-related definition performed suboptimally because keywords were not sufficiently specific. The revised definition is an improvement but requires testing in real-world scenarios. The presented methods for developing a syndromic surveillance definition are useful, not only for occupational health, but for identification of other subpopulations treated in the ED.

CIPHER: Cybersecurity Intelligent Penetration-Testing Helper for Ethical Researcher

Penetration testing, a critical component of cybersecurity, typically requires extensive time and effort to find vulnerabilities. Beginners in this field often benefit from collaborative approaches with the community or experts. To address this, we develop Cybersecurity Intelligent Penetration-testing Helper for Ethical Researchers (CIPHER), a large language model specifically trained to assist in penetration testing tasks as a chatbot. Unlike software development, penetration testing involves domain-specific knowledge that is not widely documented or easily accessible, necessitating a specialized training approach for AI language models. CIPHER was trained using over 300 high-quality write-ups of vulnerable machines, hacking techniques, and documentation of open-source penetration testing tools augmented in an expert response structure. Additionally, we introduced the Findings, Action, Reasoning, and Results (FARR) Flow augmentation, a novel method to augment penetration testing write-ups to establish a fully automated pentesting simulation benchmark tailored for large language models. This approach fills a significant gap in traditional cybersecurity Q&A benchmarks and provides a realistic and rigorous standard for evaluating LLM’s technical knowledge, reasoning capabilities, and practical utility in dynamic penetration testing scenarios. In our assessments, CIPHER achieved the best overall performance in providing accurate suggestion responses compared to other open-source penetration testing models of similar size and even larger state-of-the-art models like Llama 3 70B and Qwen1.5 72B Chat, particularly on insane difficulty machine setups. This demonstrates that the current capabilities of general large language models (LLMs) are insufficient for effectively guiding users through the penetration testing process. We also discuss the potential for improvement through scaling and the development of better benchmarks using FARR Flow augmentation results.

ENHANCING TOTAL TEMPERATURE MEASUREMENT ACCURACY: CALIBRATION PROCEDURES AND NOVEL TWO-WIRE PROBES

Abstract Total Temperature measurement with thermocouple probes suffers from errors due to heat transfer effects. Two dominant sources of errors are convection and conduction between the thermocouple, the support, and the flow. These effects can be treated in two different categories: the velocity error, created by convection from the internal flow velocity in the probe shield, and the conduction error, involving heat transfer through the wire to the shield and the probe stem due to temperature differences between each component. This paper presents a robust approach to experimentally assess and reduce both errors. An open jet test stand at the Purdue Experimental Turbine Aerothermal Laboratory is used to evaluate the effects of the velocity error at various Mach numbers. Infrared thermography measurements have been conducted to assess temperature gradients on the probe support during this calibration. With this information, it is possible to correct for conduction errors during the calibration and obtain a recovery factor that is solely dependent on upstream velocity. Recovery factor of 0.94 +/−0.05 for Mach 0.2 and 0.98 +/−0.005 for Mach 0.9 are obtained Progress has been made on a two-wire thermocouple approach to address conduction errors in the testing scenario, where IR thermography is not possible. Advances in the design and manufacturing of the probe are presented, facilitating faster design iterations and implementation, and providing full control over the calibration procedures. This method can correct conduction errors within 0.2 K for Mach 0.6

Unmanned Agricultural Machine Operation System in Farmland Based on Improved Fuzzy Adaptive Priority-Driven Control Algorithm

Autonomous driving technology for agricultural machinery can maximise crop yield, reduce labour costs, and alleviate labour intensity. In response to the current low degree of automation and low tracking accuracy of driving paths in agricultural equipment, this research proposes an unmanned agricultural machinery operating system based on an improved fuzzy adaptive PD control algorithm. Firstly, mechanical kinematic models and fuzzy adaptive control algorithms are introduced to achieve autonomous driving, and parameter settings and speed adjustments are made while considering errors. Secondly, in the autonomous driving operation system, taking a certain rice machine as an example, perception information, trajectory design, dynamic control, operation supervision, and remote control design are carried out. The experimental results show that the improved fuzzy algorithm exhibits smaller deviation results in driving path tracking, with an average error between the actual path and the expected path of less than 0.001 m. In different testing scenarios, compared with the actual control results, the maximum deviation of the control system platform in straight sections is less than 2.8 m, which is more stable. More than 95% of the lateral deviation results in the road sections are within 0.11 m. And the tracking distance error of the proposed method in the straight and curved segments is relatively small, far smaller than other comparative algorithms. The unmanned agricultural machinery operation system proposed in this study can significantly improve the efficiency and accuracy of agricultural machinery work, promote the development of intelligent and modern agricultural machinery, and provide reference value and important contributions to social and economic development as well as the progress and promotion of related technologies.

Implementation of thermal conduction energy transfer models in the Bifrost solar atmosphere MHD code

Thermal conductivity provides important contributions to the energy evolution of the upper solar atmosphere, behaving as a non-linear concentration-dependent diffusion equation. Recently, different methods have been offered as best-fit solutions to these problems in specific situations, but their effectiveness and limitations are rarely discussed. We have rigorously tested the different implementations of solving the conductivity flux, in the massively parallel magnetohydrodynamics code, Bifrost, with the aim of specifying the best scenarios for the use of each method. We compared the differences and limitations of explicit versus implicit methods, and analyse the convergence of a hyperbolic approximation. Among the tests, we used a newly derived first-order self-similar approximation to compare the efficacy of each method analytically in a 1D pure-thermal test scenario. We find that although the hyperbolic approximation proves the most accurate and the fastest to compute in long-running simulations, there is no optimal method to calculate the mid-term conductivity with both accuracy and efficiency. We also find that the solution of this approximation is sensitive to the initial conditions, and can lead to faster convergence if used correctly. Hyper-diffusivity is particularly useful in aiding the methods to perform optimally. We discuss recommendations for the use of each method within more complex simulations, whilst acknowledging the areas where there are opportunities for new methods to be developed.

Improving the industrial defect recognition in radiographic testing by pre-training on medical radiographs

Deep learning methodologies have gained substantial traction for defect recognition in industrial radiographic testing including welds, castings and other fields. Regardless of the deep learning utilized, it has emerged as a standard configuration to use a model pre-trained from ImageNet to accelerate convergence and enhance recognition accuracy. However, there is a significant gap between the domain of natural images and industrial radiographs, raising the question of whether there might be a superior pre-training method than relying on ImageNet pre-training. Fortunately, medical radiographs are more similar to industrial radiographs than natural images because of the same imaging method. In this paper, we initially utilize numerous medical radiographic images from CheXpert dataset to train a pre-trained CNN model. Then, we apply this model to four distinct tasks within two radiographic testing scenarios to validate its advantages and generalization capabilities. Finally, experiments on multiple datasets indicate that our method brings more benefits than ImageNet pre-training or training from scratch, with a F1 score improvement of 3.41 %–13.72 % for defect classification and a mIoU improvement of 1.05 %–6.58 % for defect segmentation. It demonstrates that pre-training from medical radiographs provides a cost-free improvement for all kinds of tasks in industrial defect recognition.

PERFORMANCE TESTING ANALYSIS OF BANDUNGTANGINAS APPLICATION WITH JMETER

The bandungtanginas.id application was created to serve the needs of data collection on maternal andchild health at every posyandu in the city of Bandung. This application is planned to be able to servearound 1,982 posyandu in the city of Bandung, around 10,000 users, who are expected to be able toaccess the bandungtanginas.id application. In the current condition of the application supportinfrastructure still has limitations in supporting the bandungtanginas.id application. Withspecifications that tend to be low, an improvement effort is needed through generating applicationusage scenarios so that the services provided can be optimally available using the existinginfrastructure. This research focuses on analyzing the application performance by using load testing.This research has four stages: determining the test's purpose, creating a test scenario, carrying outscenario testing, and analyzing the results. The analysis will be carried out with the results of loadtesting experiments based on the scenarios that have been made. Test scenarios are made to obtainexperimental data, last further analysis and conclusions will be made. Scenarios are created based onan estimate of the optimal number of users the application should serve. The test scenario will use atarget user of between 10 – 200 users who are divided in five groups of load testing scenarios. Fromthis trial, it will be seen how the performance of the bandungtanginas.id application for each scenariogroup that has been created. The result of this study showed that the best scenario forbandungtanginas.id application could serve users in the range of 30-50 users. It is obtained from theresponse time results, which show the average results are under 30 seconds according to the SLA(Service Level Agreement) standard for application users.

Advancements in smart agriculture through innovative weed management using wavelet-based convolution neural network

Smart agriculture has shifted the paradigm by integrating advanced technologies, particularly weed management. This paper introduces an innovative approach to weed control by applying a Wavelet-based Convolution Neural Network (WCNN). In the era of precision agriculture, our study explores the integration of WCNN into real-world scenarios, emphasizing its adaptability to diverse environmental conditions. Utilizing the spatial-frequency analysis features of wavelets and convolutional neural networks, the WCNN model is the most effective at finding weeds, classifying them, and managing them specifically in agricultural fields in real-time. This research contributes to the scientific discourse on smart agriculture and addresses the challenges of invasive weeds, presenting a sustainable solution for optimizing resource utilization. Our investigation includes a detailed exploration of WCNN’s adaptive learning mechanisms and dynamic adjustment to changing agricultural landscapes. The model seamlessly integrates with existing smart farming infrastructure, showcasing a substantial reduction in manual intervention and a simultaneous increase in agricultural productivity. We incorporate fog computing and resource optimization into our framework, enhancing the efficiency of onboard data processing. To evaluate the real-world efficacy of WCNN, we conducted comprehensive experiments in texture classification and image labelling using two distinct datasets: the plant seedling and soybean weed datasets. Results demonstrate the superior performance of WCNN, achieving higher accuracy in training and test scenarios with significantly fewer parameters than traditional CNNs. For the soybean weed dataset, WCNN achieved remarkable accuracy in the training (0.9970) and testing (0.9987) phases, with correspondingly low losses of 0.0109 and 0.0048. The WCNN model demonstrated high accuracy during training (0.9739) and testing (0.9902), with minimal losses of 0.0898 and 0.0239 in the plant seedling dataset.

Total Power Factor Smart Contract with Cyber Grid Guard Using Distributed Ledger Technology for Electrical Utility Grid with Customer-Owned Wind Farm

In modern electrical grids, the numbers of customer-owned distributed energy resources (DERs) have increased, and consequently, so have the numbers of points of common coupling (PCC) between the electrical grid and customer-owned DERs. The disruptive operation of and out-of-tolerance outputs from DERs, especially owned DERs, present a risk to power system operations. A common protective measure is to use relays located at the PCC to isolate poorly behaving or out-of-tolerance DERs from the grid. Ensuring the integrity of the data from these relays at the PCC is vital, and blockchain technology could enhance the security of modern electrical grids by providing an accurate means to translate operational constraints into actions/commands for relays. This study demonstrates an advanced power system application solution using distributed ledger technology (DLT) with smart contracts to manage the relay operation at the PCC. The smart contract defines the allowable total power factor (TPF) of the DER output, and the terms of the smart contract are implemented using DLT with a Cyber Grid Guard (CGG) system for a customer-owned DER (wind farm). This article presents flowcharts for the TPF smart contract implemented by the CGG using DLT. The test scenarios were implemented using a real-time simulator containing a CGG system and relay in-the-loop. The data collected from the CGG system were used to execute the TPF smart contract. The desired TPF limits on the grid-side were between +0.9 and +1.0, and the operation of the breakers in the electrical grid and DER sides was controlled by the relay consistent with the provisions of the smart contract. The events from the real-time simulator, CGG, and relay showed a successful implementation of the TPF smart contract with CGG using DLT, proving the efficacy of this approach in general for implementing electrical grid applications for utilities with connections to customer-owned DERs.

Comparison of Criteria Weight Determination Using MEREC and CRITIC Methods in Choosing The Best Student Accommodation with the MOORA Method Case Study: Coventry University

One of the challenges faced by IISMA Awardees and students in general in Coventry University is choosing a comfortable place to live. Although various student accommodations are provided, differences in facilities and considerations from other parties such as parents and friends make the selection process complicated. This study develops a decision support system to help students choose student accommodation objectively without any intervention from others and provides a comparison of the use of different combinations of methods as additional guidance in the decision-making process. Two methods, Method Based on the Removal Effects of Criteria (MEREC) and Criteria Importance Through Intercriteria Correlation (CRITIC), are used to weight the criteria. The Multi-Objective Optimization (MOORA) method is used to determine the best alternative after the weight calculation is known. The results using a combination of the MEREC-MOORA method and a combination of the CRITIC-MOORA method place Alternative 5 (A5) in first place, while the remaining alternatives show a similar ranking order. In this study, scenario testing was also carried out by deleting and adding criteria and alternatives which then provided ranking results with a positive correlation even though different combinations of methods were used in determining the ranking.

Construction of virtual test scenarios for intelligent car-pedestrian interaction

Aiming at the needs of intelligent car and pedestrian interaction testing under urban conditions, this paper proposes a scenario generation method that comprehensively considers the frequency of scenarios in the real world and the degree of challenge to human-vehicle interaction performance. First, according to the key features of intelligent car and pedestrian interaction, the original scene data of pedestrians crossing the road is extracted from the natural driving dataset; then, according to the needs of accelerated testing, a key scenario extraction method based on importance sampling theory is designed to extract and construct important scenarios for intelligent car-pedestrian interaction testing from the original scenes; finally, by comparing the data distribution of important scenarios with original scenarios, it is shown that this method can effectively screen out scenarios that may challenge safety performance during driving, thereby realizing accelerated testing, while taking into account the statistical characteristics of the test scenarios.

Use of neural networks for safe collision avoidance for groups of autonomous vessels

This article analyzes and compares the effectiveness of two algorithms for collision avoidance in groups of autonomous vessels: one based on geometric analysis of vessel approach and cost function minimization for calculating a safe maneuver (traditional algorithm), and an algorithm utilizing a neural network. Both algorithms assume external control of a group of vessels within a specific area using a vessel traffic management system and coordinated maneuvering of dangerously approaching vessels. Descriptions of these algorithms are provided along with their simplified block diagrams; to solve the task of safely maneuvering a group of vessels, a sequential analysis of all possible vessel pairs in the group and changes in their courses is proposed. The process of creating three test datasets is described, two of which were generated using a program and included 100 scenarios each, while the third was manually composed and included 30 scenarios for different vessel group approach variations. During the testing of the neural network algorithm, two neural networks trained to predict safe courses for vessel pairs were utilized. The neural network used in the algorithm, trained on 743671 samples, allowed the processing of test vessel approach scenarios with an accuracy comparable to the traditional algorithm. Depending on the number of dangerously approaching vessels in the area, the neural network algorithm processed test scenarios 2–14 times faster than the traditional algorithm. The paper highlights the limitations of the described algorithms and outlines planned improvements for subsequent research, including the optimization of the safe maneuver selection methodology and further training of the neural network on larger data volumes.

Instance Space Analysis of Testing of Autonomous Vehicles in Critical Scenarios

Before being deployed on roads, Autonomous Vehicles (AVs) must undergo comprehensive testing. Safety-critical situations, however, are infrequent in usual driving conditions, so simulated scenarios are used to create them. A test scenario comprises static and dynamic features related to the AV and the test environment; the representation of these features is complex and makes testing a heavy process. A test scenario is effective if it identifies incorrect behaviors of the AV. In this article, we present a technique for identifying key features of test scenarios associated with their effectiveness using Instance Space Analysis (ISA). ISA generates a ( \(2D\) ) representation of test scenarios and their features. This visualization helps to identify combinations of features that make a test scenario effective. We present a graphical representation of each feature that helps identify how well each testing technique explores the search space. While identifying key features is a primary goal, this study specifically seeks to determine the critical features that differentiate the performance of algorithms. Finally, we present metrics to assess the robustness of testing algorithms and the scenarios generated. Collecting essential features in combination with their values associated with effectiveness can be used for selection and prioritization of effective test cases.

Optimization of office process task allocation based on deep reinforcement learning

In office platforms, we often need to face the situation of a large number of parallel heterogeneous process tasks. This not only tests the ability of the task executors themselves, but also puts forward requirements for the performance of the collaborative scheduling system. Using the reinforcement learning method, combined with quantitative analysis such as collaborative coordination and slackness, and based on the Markov Bollinger theory, a multi-agent Bollinger model is proposed to realize the optimization scheduling system with overall process coordination and maximum completion time as the optimization objectives, thereby improving the overall execution efficiency. Taking the real business system process as the test scenario, under the same optimization objective, the reinforcement learning algorithm based on D3QN and DRL and the meta-heuristic algorithm based on ant colony are compared to verify the effectiveness of the proposed method.

A Developer-Oriented Framework for Assessing Power Consumption in Mobile Applications: Android Energy Smells Case Study

Currently, people spend a lot of time using their mobile devices. With such ubiquity of mobile devices in our daily life, battery capacity and quality are of utmost importance. Running software applications (called apps) are one of the major factors influencing the power consumption in mobile devices. In order to meet user needs, mobile apps are becoming inherently complex and resource greedy. Therefore, fulfilling these requirements at the pace imposed by the market may degrade software construction quality and insert so-called energy code smells: bad patterns in the source code of an app that indicate a deeper problem and adversely affect power consumption. This work proposes a developer-oriented framework for identifying and fixing patterns via analyzing different application code flavors in a user-driven test scenario. A special app was designed in order to validate the Android implementation of the proposed methodology. The study results have shown significant improvement regarding energy efficiency after correcting one or more energy code smells, with a 4 to 30 percent decrease in battery drain. Additionally, the power consumption signature term is defined in the context of mobile applications. This paper presents a developer-oriented framework for assessing power consumption in mobile applications. Our key contributions include identifying significant energy code smells, demonstrating their impact on power consumption, and providing a toolset for developers to improve energy efficiency.

Selective testing and its effect on false discovery rate controlling procedures under discrete framework

The support of the p-value corresponding to a discrete test-statistic is a finite isolated subset of the unit interval. In differential gene expression studies with RNA-sequence datasets, multiple discrete tests are performed. The objective of such studies are to identify differentially expressed genes with a control over a measure of false discoveries. We suggest to remove hypotheses corresponding to the p-values having minimum support-point greater than the threshold of that overall measure. We argue this suggestion and present methods of filtering the set of hypotheses for two useful testing scenarios. Simulation experiments show gain in power of the multiple testing procedures when applied to the filtered set of hypotheses. We also demonstrate the method with two real datasets, related to the study of HIV and mammary gland.