Real Car Research Articles

A reinforcement learning model for autonomous vehicles with realistic car simulation in urban using unity

Abstract Simulator training with reinforcement learning (RL) for autonomous vehicles (AVs) offers advantages over supervised learning. However, transferring the learned behaviours to the real world is a challenging task due to the inconsistencies between the data captured by the vehicle's sensors in the simulated environment and the real world. Additionally, some of the sensors that the AVs rely on are sensitive to weather and lighting conditions. Our proposed model addresses the challenges of sensor data disparity and environmental variation. It utilizes three sensing components which are radio detection and ranging (RADAR), inertial measurement units (IMUs), and global positioning systems (GPSs) to overcome the addressed drawbacks. The proposed model incorporates a carefully designed reward system and prioritizes computational efficiency by using fewer number of sensors and ensures safe and efficient driving. The chosen sensors enable easier knowledge transfer from the simulator to the real-world due to their consistent data representation. The model leverages the Unity engine and ML agent to train AVs for both urban and highway environments. The experimental results show that our suggested model effectively trained AVs to navigate through complex urban areas without collisions while keeping them in their lanes. The demonstration video is provided in the following link: https://youtu.be/YCOjli7lrCM

A single center retrospective study on real world CAR T-cell therapy: focus on early hematological toxicity.

Patients with relapsed/refractory diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma (MCL) and multiple myeloma have poor outcomes. CAR-T completely changed the landscape of therapy options, improving not only response rates but also survival outcomes. Hematological toxicity after chimeric antigen receptor therapy (CAR-T) is of increasing interest, being a recognized prognostic factor in this setting. We report our experience with early hematological toxicity after CAR-T therapy and point some important aspects regarding the Hematotox-Score. We identified a strong negative correlation between Hematotox-Score and platelet count at first day of cytokine release syndrome (CRS). Hematotox-Score was predictive of hemoglobin levels at day 28 after CAR-T. Ferritin remained high after 28 days post CAR-T in patients with high Hematotox-Score. Hematotox-Score did not associate with mortality in our cohort. We did not find any significant association between the hematological parameters (hemoglobin, platelets, and neutrophil counts), ferritin, LDH at first day of CRS and mortality. In conclusion, we demonstrate that Hematotox-Score is predictive of early hematological toxicity after CAR-T. Although, patients with higher degree of hematological toxicities have poorer survival outcomes, Hematotox-Score lacks predictive potential, probably due to its limitations. Further development of hematological scores predicting survival outcome in the context of CAR-T are needed.

Improving Freight Train Wheel Monitoring with Smart Sensors

Onboard monitoring of freight car axleboxes enhances safety, reduces maintenance costs, and improves track conditions by preventing secondary damage. Installing wireless sensors on freight cars without a nearby power source should be cost-effective, given the large quantities involved. To address this, a new wireless smart sensor node has been deployed. The sensor automatically recognizes stable operating conditions, detects wheel rotational speed from vibrations, performs real-time condition monitoring, and transmits the results to the cloud. This study outlines the smart sensor concept and the pilot field test conducted with real freight cars. The results demonstrate the ability to estimate wheel rotational speed from vibrations and the potential for detecting wheel out-of-roundness (OOR) using a newly developed condition indicator for low-power real-time operations.

Effect Analysis of Different Use Ways of a Q10 Child Seat

In this paper, the injury situation of Q10 child dummy was optimized and analyzed by analyzing the impact of different ways of using a certain child seat ,and analyzing the crash test data of a real car based on LS-DYNA crash simulation analysis software after the vehicle benchmarking, by elaborating on the protection of child occupants in different ways of using this type of child seat, so as to provide a reference for the development of child restraint system of new cars. The results show that: Regardless of whether this type of child seat has an optional seat back, adding seat belt Secure Guard and not using ISOFIX fixing, the dummies' injuries show a decreasing trend. This study provides reference for the selection and use of this type of child seat in dynamic test.

Comparison of different methods for categorizing the individual susceptibility to motion sickness – A carsickness study

In studies aimed at developing avoidance strategies to reduce motion sickness (kinetosis) in autonomous vehicles, failing to account for the wide variability in individual kinetosis susceptibility can lead to inaccuracies and disregard effective countermeasures. Three methods for assessing individual susceptibility to carsickness – two questionnaires focusing on kinetosis experiences and a kinetosis-provoking lab test – were compared with the development of kinetosis during real car driving tests. Questions about car-specific kinetosis-provoking situations (MS-C) exhibit stronger correlations with kinetosis in car experiments compared to the commonly used questions about kinetosis experiences across different types of transportation (MS-VD). While lab-based testing remains highly reliable, especially considering men's tendency to underestimate their carsickness susceptibility in questionnaires, MS-C provides a valuable compromise in terms of technical and time expenses. These findings can also be used to assist passengers of autonomous driving cars in accurately assessing their sensitivity and activating customized countermeasure functions.

Can we Use Smart Phone on a Moving Vehicle Without Worrying About Carsickness? Developing an Effective Motion Cue APP with Driving Simulator and Real Vehicle Experiments

The prevalence of motion sickness among passengers using personal electronic devices, such as smartphones, during vehicle journeys has become a growing concern. This issue is expected to intensify with the increasing adoption of assistant or automated driving functions, which may lead to non-driving tasks (NDT) being performed by all on-board passengers, including the user in the “driver” seat during conditionally or fully automated driving modes. This trend presents challenges related to motion sickness, particularly in terms of specific performance requirements for non-driving tasks. In response to the need to alleviate passenger motion sickness, we have developed an easy-to-understand animation cue app that can be conveniently implemented on smartphones. The motion cue conveys information about vehicle accelerations, including their directions and magnitudes, using the metaphors of traffic signal colors and backward-moving lane lines, either in straight or curved lane driving. Following several rounds of improvements based on moving-base simulator and real car experiments, finally a successful cue design was found, which could significantly alleviate motion sickness of passengers while engaging in NDT, with minimal impact on their NDT performances. However, the study also revealed limitations to the sickness-alleviating capability of our motion cue design, including potential lack of universal acceptance among different users and reduced effectiveness in severely uncomfortable driving conditions. This work may provide valuable insights for further visual cue improvements that can contribute in future carsickness-proof vehicles.

Data-efficient multimodal human action recognition for proactive human–robot collaborative assembly: A cross-domain few-shot learning approach

With the recent vision of Industry 5.0, the cognitive capability of robots plays a crucial role in advancing proactive human–robot collaborative assembly. As a basis of the mutual empathy, the understanding of a human operator’s intention has been primarily studied through the technique of human action recognition. Existing deep learning-based methods demonstrate remarkable efficacy in handling information-rich data such as physiological measurements and videos, where the latter category represents a more natural perception input. However, deploying these methods in new unseen assembly scenarios requires first collecting abundant case-specific data. This leads to significant manual effort and poor flexibility. To deal with the issue, this paper proposes a novel cross-domain few-shot learning method for data-efficient multimodal human action recognition. A hierarchical data fusion mechanism is designed to jointly leverage the skeletons, RGB images and depth maps with complementary information. Then a temporal CrossTransformer is developed to enable the action recognition with very limited amount of data. Lightweight domain adapters are integrated to further improve the generalization with fast finetuning. Extensive experiments on a real car engine assembly case show the superior performance of proposed method over state-of-the-art regarding both accuracy and finetuning efficiency. Real-time demonstrations and ablation study further indicate the potential of early recognition, which is beneficial for the robot procedures generation in practical applications. In summary, this paper contributes to the rarely explored realm of data-efficient human action recognition for proactive human–robot collaboration.

Sim-to-Real Application of Reinforcement Learning Agents for Autonomous, Real Vehicle Drifting

Enhancing the safety of passengers by venturing beyond the limits of a human driver is one of the main ideas behind autonomous vehicles. While drifting is mostly witnessed in motorsports as an advanced driving technique, it could provide many possibilities for improving traffic safety by avoiding accidents in extreme traffic situations. The purpose of the research presented in this article is to provide a machine learning-based solution to autonomous drifting as a proof of concept for vehicle control at the limits of handling. To achieve this, reinforcement learning (RL) agents were trained for the task in a MATLAB/Simulink-based simulation environment, using the state-of-the-art Soft Actor–Critic (SAC) algorithm. The trained agents were tested in reality at the ZalaZONE proving ground on a series production sports car with zero-shot transfer. Based on the test results, the simulation environment was improved through domain randomization, until the agent could perform the task both in simulation and in reality on a real test car.

Model-Based 3D Gaze Estimation Using a TOF Camera.

Among the numerous gaze-estimation methods currently available, appearance-based methods predominantly use RGB images as input and employ convolutional neural networks (CNNs) to detect facial images to regressively obtain gaze angles or gaze points. Model-based methods require high-resolution images to obtain a clear eyeball geometric model. These methods face significant challenges in outdoor environments and practical application scenarios. This paper proposes a model-based gaze-estimation algorithm using a low-resolution 3D TOF camera. This study uses infrared images instead of RGB images as input to overcome the impact of varying illumination intensity in the environment on gaze estimation. We utilized a trained YOLOv8 neural network model to detect eye landmarks in captured facial images. Combined with the depth map from a time-of-flight (TOF) camera, we calculated the 3D coordinates of the canthus points of a single eye of the subject. Based on this, we fitted a 3D geometric model of the eyeball to determine the subject's gaze angle. Experimental validation showed that our method achieved a root mean square error of 6.03° and 4.83° in the horizontal and vertical directions, respectively, for the detection of the subject's gaze angle. We also tested the proposed method in a real car driving environment, achieving stable driver gaze detection at various locations inside the car, such as the dashboard, driver mirror, and the in-vehicle screen.

Machine Learning to Infer a Health State Using Biomedical Signals — Detection of Hypoglycemia in People with Diabetes while Driving Real Cars

Machine Learning to Infer a Health State Using Biomedical Signals — Detection of Hypoglycemia in People with Diabetes while Driving Real Cars

Vehicle pitch dynamics control using in-wheel motors

Implementation of in-wheel motors brings numerous advantages: decreased overall vehicle weight and manufacture cost, available accurate torque control for better road handling and powertrain high energy efficiency. Even so, when the motor is placed inside the wheel, the unsprung mass of the vehicle increases, leading to a less comfortable ride. This paper puts forward and studies a control system that takes advantage of the vertical component of the driving force provided by an in-wheel motor to enhance ride comfort and handling dynamics. The proportional integral (PI) controller was proposed to minimize the pitch acceleration response of the vehicle. Controller’s gains were optimized using the gradient descent method with sequential quadratic programing algorithm. A co-simulation approach and a dynamic model of the vehicle were presented. Data regarding the vehicle’s measurements was taken from a real car with in-wheel motors. The dynamic model of the car was validated based on the results derived during the road test experiment. Assessing the controller’s effectiveness, a co-simulation approach between the dynamic model and the control system was established. An acceleration test in a straight line showed a 25% increase in ride comfort as measured by the RMS pitch angular vibrations of the sprung mass. The control is attained with no increase in the tire load and the vehicle’s longitudinal dynamics.

Effectiveness and User Perception of an In-Vehicle Voice Warning for Hypoglycemia: Development and Feasibility Trial.

Hypoglycemia is a frequent and acute complication in type 1 diabetes mellitus (T1DM) and is associated with a higher risk of car mishaps. Currently, hypoglycemia can be detected and signaled through flash glucose monitoring or continuous glucose monitoring devices, which require manual and visual interaction, thereby removing the focus of attention from the driving task. Hypoglycemia causes a decrease in attention, thereby challenging the safety of using such devices behind the wheel. Here, we present an investigation of a hands-free technology-a voice warning that can potentially be delivered via an in-vehicle voice assistant. This study aims to investigate the feasibility of an in-vehicle voice warning for hypoglycemia, evaluating both its effectiveness and user perception. We designed a voice warning and evaluated it in 3 studies. In all studies, participants received a voice warning while driving. Study 0 (n=10) assessed the feasibility of using a voice warning with healthy participants driving in a simulator. Study 1 (n=18) assessed the voice warning in participants with T1DM. Study 2 (n=20) assessed the voice warning in participants with T1DM undergoing hypoglycemia while driving in a real car. We measured participants' self-reported perception of the voice warning (with a user experience scale in study 0 and with acceptance, alliance, and trust scales in studies 1 and 2) and compliance behavior (whether they stopped the car and reaction time). In addition, we assessed technology affinity and collected the participants' verbal feedback. Technology affinity was similar across studies and approximately 70% of the maximal value. Perception measure of the voice warning was approximately 62% to 78% in the simulated driving and 34% to 56% in real-world driving. Perception correlated with technology affinity on specific constructs (eg, Affinity for Technology Interaction score and intention to use, optimism and performance expectancy, behavioral intention, Session Alliance Inventory score, innovativeness and hedonic motivation, and negative correlations between discomfort and behavioral intention and discomfort and competence trust; all P<.05). Compliance was 100% in all studies, whereas reaction time was higher in study 1 (mean 23, SD 5.2 seconds) than in study 0 (mean 12.6, SD 5.7 seconds) and study 2 (mean 14.6, SD 4.3 seconds). Finally, verbal feedback showed that the participants preferred the voice warning to be less verbose and interactive. This is the first study to investigate the feasibility of an in-vehicle voice warning for hypoglycemia. Drivers find such an implementation useful and effective in a simulated environment, but improvements are needed in the real-world driving context. This study is a kickoff for the use of in-vehicle voice assistants for digital health interventions.

Real Time Car Model and Plate Detection System by Using Deep Learning Architectures

Real Time Car Model and Plate Detection System by Using Deep Learning Architectures

ИССЛЕДОВАНИЕ ПРОЦЕССА ИЗНАШИВАНИЯ ШЕЕК КОЛЕНЧАТОГО ВАЛА, ВОССТАНОВЛЕННЫХ УСТАНОВКОЙ СТАЛЬНЫХ ПОЛУКОЛЬЦЕВЫХ НАКЛАДОК

Installation of steel semi-ring linings on the crankshaft journals is a rational technological solution that can significantly reduce the thermal deformation of the crankshaft in the process of its restoration compared to building-up welding or tape welding. (Research purpose) The research purpose is studying the wear rate in crankshaft journals restored by installing steel semi-ring linings. (Materials and methods) The wear rate in the restored crankshaft journals has been theoretically calculated using I.V. Kragelsky’s method. The diesel engine crankshaft in a Volkswagen Transporter car, which was in operation, has been taken as an experimental sample. The semi-ring linings have been installed on the crankshaft journals by butt welding to each other. The main criterion for the experimental evaluation is the wear rate of the restored journals under real operating conditions. When measuring wear, both the profilogram method and the micrometric method have been used for measuring the diameters of the crankshaft journals before and after a certain car run. (Results and discussion) The values of the main journal wear during real car run are comparable to the theoretical calculation results. The wear degree of the connecting rod journals exceeds the wear of the main ones, which can be explained by absence of heat treatment of the half rings in the connecting rod journals when restoring the shaft. The actual values of ellipsity as well as taper rates of the journals are within the acceptable values. The wear values obtained by measuring roughness before and after the test period are comparable to those obtained by micrometric measurement. (Conclusions) The wear rate of both connecting rod and main crankshaft journals is comparable to the values determined by theoretical calculation, which makes it possible to conclude that the experimental research has a positive result. The wear rates determined both experimentally and theoretically are within the limits established by earlier studies in the field of tribology, which indicates the reliability of the experimental results. Thus, the applicability of the method of installing steel semi-ring linings when restoring crankshafts of automotive engines has been confirmed.

On the effectiveness of Reynolds-averaged and subgrid scale models in predicting flows inside car cabins

The aim of the present study is to analyze the performances of unsteady Reynolds-averaged Navier–Stokes (URANS) and large eddy simulation (LES) approaches in predicting the airflow patterns inside car cabins and to give insight in the design of computational fluid dynamics simulations of a real car cabin. For this purpose, one eddy viscosity-based turbulence model (shear stress transport k–ω) and two subgrid scale models (wall-adapting local eddy-viscosity and dynamic kinetic energy) were tested, and numerical results were compared with particle image velocimetry measurements carried out on a commercial car. The URANS model exhibited great accuracy in predicting the mean flow behavior and was appreciably outperformed by the LES models only far from the inlet sections. For this reason, it was deemed suitable for conducting further analyses, aimed at characterizing the airflow patterns in winter and summer conditions and performing a thermal comfort analysis. The thermal regime was found to have a very little effect on the air flow patterns, once the quasi-steady state regime is achieved; in fact, both in winter and in summer, the temperature field is fairly uniform within the car cabin, making the contribution of buoyancy negligible and velocity fields to be very similar in the two seasons. Findings also reveal that thermal comfort sensation can be different for passengers sharing the same car but sitting on different seats; this aspect should be considered when designing and operating the ventilation system, since the minimum comfort requirements should be met for all the occupants.

Digital measurement of hands-on performance? Ecological validation of a computer-based assessment of automotive repair skills

We investigated the ecological validity of performance measures from a computer-based assessment tool that utilises scripted video vignettes. The intended purpose of this tool is to assess the maintenance and repair skills of automotive technician apprentices, complementing traditional hands-on assessment formats from the German journeymen’s exams. We hypothesise that the ability to correctly judge repair actions shown in videos is a good predictor of the ability to perform corresponding actions in hands-on scenarios. Apprentices in the third year of vocational training carried out repairs on real cars or car systems, while experts rated their performance. After this, they worked on our computer-based tests, which utilise videos of very similar repairs. The correlation between video judgement and hands-on performance was lower than expected for most repair actions as well as for overall scores, indicating insufficient ecological validity of the test score interpretations. However, the findings are promising for developing future tests, as the results for some repair actions indicate it is generally possible to develop ecologically valid video-based items focusing on hands-on skills. We discuss the results in the light of a validation framework that combines validity evidence from different sources for the same assessment tool. Finally, we hope our findings contribute to a broader discussion about the psychometric quality of exams.

A wide dynamic range diamond quantum sensor as an electric vehicle battery monitor.

To demonstrate the application capability of the diamond quantum sensor as an electric vehicle (EV) battery monitor, we (i) investigated the measurable current in a real car noise level and (ii) compared the linearity with conventional sensors. Consequently, (i) we could measure a 20 mA current pulse even under an external magnetic field of 80 µT, which is larger than that of 50 µT around the EV battery module in a real car during driving. The 20 mA pulse measurement corresponds to the EV battery state of charge estimation accuracy of 0.2% in the standard driving pattern, which is smaller than the present level of 10%. (ii) The linearity degradation seen in the Hall sensor near the upper limit of the measurement range was not seen in the diamond sensor. Although the Hall sensor and the shunt resistor showed linearity degradation in the current range of several tens of amperes or less, the degradation was smaller for the diamond sensor. The transverse magnetic field effect in the diamond sensor on the linearity was estimated to be less than 0.01% for a several-degree misalignment of the sensor surface to the magnetic field direction and under a 340 A current. This article is part of the Theo Murphy meeting issue 'Diamond for quantum applications'.

Car drag coefficient prediction using long–short term memory neural network and LASSO

Although the methods of car's shape designing have undergone significant changes, the drag coefficient has consistently act as a crucial indicator of a car's fuel efficiency and environmental impact. Nowadays, the analysis of the drag coefficient for car bodies relies heavily on CAE, which is a time-consuming process. In this paper, a Long Short-Term Memory (LSTM) prediction model which combines Deep Learning and CAD has been established to estimate the drag coefficient. Feature extraction of the Car Side Silhouettes is carried out based on Bezier curve car body contours, and Lasso regression is employed to solve overfitting problems caused by the insufficient amount of data. The prediction average deviation is 1.1% for drag coefficient during 0.28 to 0.35, and 5% for a real car profile. The results showed that the method provides better performance than the existing method, and improves accuracy of prediction as well.

Exploring the Evolution of Autonomous Vehicle Acceptance through Hands-On Demonstrations

This article delves into the acceptance of autonomous driving within society and its implications for the automotive insurance sector. The research encompasses two different studies conducted with meticulous analysis. The first study involves over 600 participants involved with the automotive industry who have not yet had the opportunity to experience autonomous driving technology. It primarily centers on the adaptation of insurance products to align with the imminent implementation of this technology. The second study is directed at individuals who have had the opportunity to test an autonomous driving platform first-hand. Specifically, it examines users’ experiences after conducting test drives on public roads using an autonomous research platform jointly developed by MAPFRE, Universidad Carlos III de Madrid, and Universidad Politécnica de Madrid. The study conducted demonstrates that the user acceptance of autonomous driving technology significantly increases after firsthand experience with a real autonomous car. This finding underscores the importance of bringing autonomous driving technology closer to end-users in order to improve societal perception. Furthermore, the results provide valuable insights for industry stakeholders seeking to navigate the market as autonomous driving technology slowly becomes an integral part of commercial vehicles. The findings reveal that a substantial majority (96% of the surveyed individuals) believe that autonomous vehicles will still require insurance. Additionally, 90% of respondents express the opinion that policies for autonomous vehicles should be as affordable or even cheaper than those for traditional vehicles. This suggests that people may not be fully aware of the significant costs associated with the systems enabling autonomous driving when considering their insurance needs, which puts the spotlight back on the importance of bringing this technology closer to the general public.

Optimal Planning Strategy for Reconfigurable Electric Vehicle Chargers in Car Parks

A conventional electric vehicle charger (EVC) charges only one EV concurrently. This leads to underutilization whenever the charging power is less than the EVC-rated capacity. Consequently, the cost-effectiveness of conventional EVCs is limited. Reconfigurable EVCs (REVCs) are a new technology that overcomes underutilization by allowing multiple EVs to be charged concurrently. This brings a cost-effective charging solution, especially in large car parks requiring numerous chargers. Therefore, this paper proposes an optimal planning strategy for car parks deploying REVCs. The proposed planning strategy involves three stages. An optimization model is developed for each stage of the proposed planning strategy. The first stage determines the optimal power rating of power modules inside each REVC, and the second stage determines the optimal number and configuration of REVCs, followed by determining the optimal operation plan for EV car parks in the third stage. To demonstrate the effectiveness of the proposed optimal planning strategy, a comprehensive case study is undertaken using realistic car parking scenarios with 400 parking spaces, electricity tariffs, and grid infrastructure costs. Compared to deploying other conventional EVCs, the results convincingly indicate that the proposed optimal planning strategy significantly reduces the total cost of investment and operation while satisfying charging demands.