Search Results

Autonomous driving is a hot topic in the automotive domain, and there is an increasing need to prove its reliability. They use machine learning techniques, which are themselves stochastic techniques based on some kind of statistical inference. The occurrence of incorrect decisions is part of this approach and often not directly related to correctable errors. The quality of the systems is indicated by statistical key figures such as accuracy and precision. Numerous driving tests and simulations in simulators are extensively used to provide evidence. However, the basis of all descriptive statistics is a random selection from a probability space. The difficulty in testing or constructing the training and test data set is that this probability space is usually not well defined. To systematically address this shortcoming, ontologies have been and are being developed to capture the various concepts and properties of the operational design domain.

Though modal analysis is a common tool to evaluate the dynamic properties of a structure, there are still many individual decisions to be made during the process which are often based on experience and make it difficult for occasional users to gain reliable and correct results. One of those experience-based choices is the correct number and placement of reference points. This decision is especially important, because it must be made right in the beginning of the process and a wrong choice is only noticeable in the very end of the process. Picking the wrong reference points could result in incomplete modal analysis outcomes, as it might make certain modes undetectable, compounded by the user's lack of awareness about these missing modes. In the paper an innovative approach will be presented to choose the minimal number of mandatory reference points and their placement.

The paper presents a theoretical framework for the detection and first-level preliminary identification of potential defects on aero-structure components while employing ultrasonic guided wave based structural health monitoring strategies, systems and tools. In particular, we focus our study on ground inspection using laser-Doppler scan of surface velocity field, which can also be partly reconstructed or monitored using point sensors and actuators on-board structurally integrated. Using direct wave field data, we first question the detectability of potential defects of unknown location, size, and detailed features. Defects could be manufacturing defects or variations, which may be acceptable from design and qualification standpoint; however, those may cause significant background signal artifacts in differentiating structure progressive damage or sudden failure like impact-induced damage and fracture.

The evaluation of aircraft characteristics through flight test maneuvers is fundamental to aviation safety and understanding flight attributes. This research project proposes a comprehensive methodology to detect and analyze aircraft maneuvers using full flight data, combining signal processing and machine learning techniques. Leveraging the Wavelet Transform, we unveil intricate temporal details within flight data, uncovering critical time-frequency insights essential for aviation safety. The integration of Long Short-Term Memory (LSTM) models enhances our ability to capture temporal dependencies, surpassing the capabilities of machine learning in isolation. These extracted maneuvers not only aid in safety but also find practical applications in system identification, air-data calibration, and performance analysis, significantly reducing pre-processing time for analysts.

The present study discusses about the determination of the Seal drag force in the application where elastomeric seal is used with metallic interface in the presence of different fluids. An analytical model was constructed to predict the seal drag force and experimental test was performed to check the fidelity of the analytical model. A Design of Experiment (DoE) was utilized to perform experimental test considering different factors affecting the Seal drag force. Statistical tools such as Test for Equal Variances and One way Analysis of Variance (ANOVA) were used to draw inferences for population based on samples tested in the DoE test. It was observed that Glycol based fluids lead to lubricant wash off resulting into increased seal drag force. Additionally, non-lubricated seals tend to show higher seal drag force as compared to lubricated seals. Keywords: Seal Drag, DoE, ANOVA

The objectives of this study were to provide insights on how injury risk is influenced by occupant demographics such as sex, age, and size; and to quantify differences within the context of commonly-occurring real-world crashes. The analyses were confined to either single-event collisions or collisions that were judged to be well-defined based on the absence of any significant secondary impacts. These analyses, including both logistic regression and descriptive statistics, were conducted using the Crash Investigation Sampling System for calendar years 2017 to 2021. In the case of occupant sex, the findings agree with those of many recent investigations that have attempted to quantify the circumstances in which females show elevated rates of injury relative to their male counterparts given the same level bodily insult. This study, like others, provides evidence of certain female-specific injuries.

Verification and validation (V&V) of autonomous vehicles (AVs) is a challenging task. AVs must be thoroughly tested, to ensure their safe functionality in complex traffic situations including rare but safety-relevant events. Furthermore, AVs must mitigate risks and hazards that result from functional insufficiencies, as described in the Safety of the Intended Functionality (SOTIF) standard. SOTIF analysis includes iterative identification of driving scenarios that are not only unsafe, but also unknown. However, identifying SOTIF’s unknown-unsafe scenarios is an open challenge. In this paper we proposed a systematic optimization-based approach for identification of unknown-unsafe scenarios. The proposed approach consists of three main steps including data collection, feature extraction and optimization towards unknown unsafe scenarios.

As a key technology of intelligent transportation system, vehicle type recognition plays an important role in ensuring traffic safety,optimizing traffic management and improving traffic efficiency, which provides strong support for the development of modern society and the intelligent construction of traffic system. Aiming at the problems of large number of parameters, low detection efficiency and poor real-time performance in existing vehicle type recognition algorithms, this paper proposes an improved vehicle type recognition algorithm based on YOLOv5. Firstly, the lightweight network model MobileNet-V3 is used to replace the backbone feature extraction network CSPDarknet53 of the YOLOv5 model. The parameter quantity and computational complexity of the model are greatly reduced by replacing the standard convolution with the depthwise separable convolution, and enabled the model to maintain higher accuracy while having faster reasoning speed.

This paper demonstrates a data-driven methodology for the system-level design of high-power traction motor drives in modern battery electric vehicles. With the immense growth of battery electric vehicles in this transformative decade, the expected time to develop and market these powertrain components is becoming significantly shorter than for internal combustion engines. This rising demand is further complicated due to more stringent cost, efficiency and power density targets set by the U.S. Department of Energy. Hence, a system-level perspective is maintained in this data-driven methodology to identify the design requirements for traction motor drives by relying on a dynamic vehicle simulation toolchain and various drive cycles (e.g., EPA MCT, WLTC, US06, etc.). The proposed data-driven approach can be used across different battery electric vehicle platforms including passenger and commercial types.

Water content estimation is a key problem for studying the PEM fuel cell. When several hundred fuel cells are connected in serial and their active surface area is enlarged for sufficient power, the difference between cells becomes significant with respect to voltage and water content. The voltage of each cell is measurable by the cell voltage monitor (CVM) while it is difficult to estimate water content of the individual. Resistance of the polymer electrolyte membrane is monotonically related to its water content, so that the new online high frequency resistance (HFR) measurement technique is investigated to identify the uniformity of water content between cells and analyze its sensitivity to operating conditions in this paper. Firstly, the accuracy of the proposed technique is experimentally validated to be comparable to that of a commercialized electrochemical impedance spectroscopy (EIS) measurement equipment.

The accurate prediction of engine performance maps can guide data-driven optimization of engine technologies to control fuel use and associated emissions. However, engine operational maps are scarcely reported in literature and often have missing data. Assessment of missing-data resilient algorithms in the context of engine data prediction could enable better processing of real-world driving cycles, where missing data is a more pervasive phenomenon. The goal of this study is, therefore, to determine the most effective technique to deal with missing data and employ it in prediction of engine performance characteristics. We assess the performance of two machine learning approaches, namely Artificial Neural Networks (ANNs) and the extreme tree boosting algorithm (XGBoost), in handling missing data.

As a key tool to maintain urban cleanliness and improve the road environment, road cleaning vehicles play an important role in improving the quality of life of residents. However, the traditional road cleaning vehicle requires the driver to monitor the situation of road garbage at all times and manually operate the cleaning process, resulting in an increase in the driver 's work intensity. To solve this problem, this paper proposes a road garbage recognition algorithm based on improved YOLOv5, which aims to reduce labor consumption and improve the efficiency of road cleaning. Firstly, the lightweight network MobileNet-V3 is used to replace the backbone feature extraction network of the YOLOv5 model. The number of parameters and computational complexity of the model are greatly reduced by replacing the standard convolution with the deep separable convolution, which enabled the model to have faster reasoning speed while maintaining higher accuracy.

In the process of designing the aerodynamic kit for Formula SAE racing cars, there is a lot of repetitive work and low efficiency in optimizing parameters such as wing angle of attack and chord length. Moreover, the optimization of these parameters in past designs heavily relied on design experience and it's difficult to achieve the optimal solution through theoretical calculations. By establishing a parametric model in CAD software and integrating it with CFD software, we can automatically modify model parameters, run a large number of simulations, and analyze the simulation results using statistical methods. After multiple iterations, we achieve fully automatic parameter optimization and obtain higher negative lift. At the same time, the simulation process is optimized, and simulations are run based on GPUs, resulting in a significant increase in simulation speed compared to the original.

The proposed Euro 7 regulation includes On Board Monitoring, or OBM, to continuously monitor vehicles for emission exceedances. OBM relies on feedback from existing or additional sensors to identify high emitting vehicles, which poses many challenges. Currently, sensors are not commercially available for all emissions constituents, and the accuracy of available sensors is not capable enough for in use compliance determination. On board emissions models do not offer enough fidelity to determine in use compliance and require new complex model innovation development which will be extremely complicated to implement on board the vehicle. The stack up of multi-component deterioration leading to an emissions exceedance is infeasible to detect using available sensors and models.

Understanding driving behavior is crucial for enhancing traffic safety. While previous studies have primarily explored driving behavior using either statistical or machine learning methods, comprehensive assessments employing both methods under various driving mode are limited. In this study, we employ both machine learning and statistical approaches to model driving behavior. First, we design a comprehensive driver grading system to assess the behavior of drivers under different driving modes. Additionally, we present an extended isolation forest-based model to classify driving behavior using data without labels, saving time and effort. Results illustrate that safe driving is more consistent and stable, while aggressive driving exhibits more intensive changes. They also demonstrate that drivers can exhibit various behaviors under different modes, serving as a benchmark for further driver modeling.

Engineering design-decisions often involve many attributes which can differ in the levels of their importance to the decision maker (DM), while also exhibiting complex statistical relationships. Learning a decision-making policy which accurately represents the DM’s actions has long been the goal of decision analysts. To circumvent elicitation and modeling issues, this process is often oversimplified in how many factors are considered and how complicated the relationships considered between them are. Without these simplifications, the classical lottery-based preference elicitation is overly expensive, and the responses degrade rapidly in quality as the number of attributes increase. In this paper, we investigate the ability of deep preference machine learning to model high-dimensional decision-making policies utilizing rankings elicited from decision makers.

In the field of autonomous driving trajectory planning, it’s virtual to ensure real-time planning while guaranteeing feasibility and robustness. Current widely adopted approaches include decoupling path planning and velocity planning based on optimization method, which can’t always yield optimal solutions, especially in complex dynamic scenarios. Furthermore, search-based and sampling-based solutions encounter limitations due to their low resolution and high computational costs. This paper presents a novel spatio-temporal trajectory planning approach that integrates both search-based planning and optimization-based planning method. This approach retains the advantages of search-based method, allowing for the identification of a global optimal solution through search. To address the challenge posed by the non-convex nature of the original solution space, we introduce a spatio-temporal semantic corridor structure, which constructs a convex feasible set for the problem.

This paper compares the results from three human factors studies conducted in a motion-based simulator in 2008, 2014 and 2023, to highlight the trends in driver's response to Forward Collision Warning (FCW). The studies were motivated by the goal to develop an effective HMI (Human-Machine Interface) strategy that enables the required driver's response to FCW while minimizing the level of annoyance of the feature. All three studies evaluated driver response to a baseline-FCW and no-FCW conditions. Additionally, the 2023 study included two modified FCW chime variants: a softer FCW chime and a fading FCW chime. Sixteen (16) participants, balanced for gender and age, were tested for each group in all iterations of the studies. The participants drove in a high-fidelity simulator with a visual distraction task (number reading). After driving 15 minutes in a nighttime rural highway environment, a surprise forward collision threat arose during the distraction task.

With the rapid development of intelligent driving technology, there has been a growing interest in the driving comfort of automated vehicles. As vehicles become more automated, the role of the driver shifts from actively engaging in driving tasks to that of a passenger. Consequently, the study of the passenger experience in automated driving vehicles has emerged as a significant research area. In order to examine the impact of automatic driving on passengers' riding experience in vehicle platooning scenarios, this study conducted real vehicle experiments involving six participants. The study assessed the subjective perception scores, eye movement, and electrocardiogram (ECG) signals of passengers seated in the front passenger seat under various vehicle speeds, distances, and driving modes. The results of the statistical analysis indicate that vehicle speed has the most substantial influence on passenger perception.

Rotary Bell Atomizers are well established in the automotive industry for top coating applications. This type of atomizer allows to create a uniform coating and is characterized by high productivity. Meanwhile, the effectiveness of the process depends on many complex factors. For instance, the transfer efficiency of the paint material, which is the percentage of the paint reaching the structure surface, ranges from 60-95% depending on the application conditions. Any increase in the transfer efficiency can not only reduce energy and material costs, but also reduce the emission of harmful non-deposited paint particles and the effort to handle them. The use of accurate numerical methods in this process helps to optimize the application process, reduce the number of expensive field experiments, and shortens the development cycle of new vehicles, which ensures predictability of production costs.