Search Results

Abstract Electric power steering (EPS) system is a kind of dynamic control system for vehicle steering, which can amplify the driver steering torque inputs to the vehicle to improve steering comfortable and performance, but the present EPS can’t cater to the driving habits of different people. In this paper, a personalized EPS controller is designed based on the driver behavior, which combines real-time driver behavior identification strategy with personalized assistance characteristic. Firstly, the driver behavior data acquisition system is designed and established, based on which, the input data of different kinds of drivers along with vehicle signals are collected under typical working conditions, then the identification of driver behavior online is realized using the BP neural network.

Abstract Letter to the Editor

Abstract Modeling, prediction, and evaluation of personalized driving behaviors are crucial to emerging advanced driver-assistance systems (ADAS) that require a large amount of customized driving data. However, collecting such type of data from the real world could be very costly and sometimes unrealistic. To address this need, several high-definition game engine-based simulators have been developed. Furthermore, the computational load for cooperative automated driving systems (CADS) with a decent size may be much beyond the capability of a standalone (edge) computer. To address all these concerns, in this study we develop a co-simulation platform integrating Unity, Simulation of Urban MObility (SUMO), and Amazon Web Services (AWS), where Unity provides realistic driving experience and simulates on-board sensors; SUMO models realistic traffic dynamics; and AWS provides serverless cloud computing power and personalized data storage.

Abstract We investigate the contribution of several variables concerning the severity of accidents involving vehicle occupant and pedestrian victims in Tunisia. In order to investigate the effect of various explanatory variables, Odds Ratio (OR) effects are considered for both serious injury accidents and fatal accidents. The empirical results are of great variety. The vehicle-occupant severity model indicates that male drivers are associated with higher severity levels as compared to female drivers. Added to that, accidents occurring in rainy conditions increase the likelihood of fatal injuries but have no significant effect on other injury severity levels. Among driver contributory factors, a driver under the influence of alcohol or drug is associated with an increased risk of sustaining fatal injuries compared to other driver contributory factors. The season factor shows that accident severity during the summer season is high.

Abstract In this work, we present a lightweight pipeline for robust behavioral cloning of a human driver using end-to-end imitation learning. The proposed pipeline was employed to train and deploy three distinct driving behavior models onto a simulated vehicle. The training phase comprised of data collection, balancing, augmentation, preprocessing, and training a neural network, following which the trained model was deployed onto the ego vehicle to predict steering commands based on the feed from an onboard camera. A novel coupled control law was formulated to generate longitudinal control commands on the go based on the predicted steering angle and other parameters such as the actual speed of the ego vehicle and the prescribed constraints for speed and steering. We analyzed the computational efficiency of the pipeline and evaluated the robustness of the trained models through exhaustive experimentation during the deployment phase.

Abstract Electric vehicles (EVs) are tested for their driving performance, energy consumption (EC), and electric range in a chassis dynamometer laboratory as per the test procedure of the Automotive Industry Standards (AIS) in India. However, a laboratory test is different from a real-world test. In this study, the test vehicle is an experimental EV conversion car developed for the study. The EV uses a prototype conversion kit consisting of an indigenously built three-phase induction motor and Lithium-ion battery (LiB) pack. The study reports a detailed discussion on the real-world as well as laboratory tests conducted for the EV. It provides an intuition of the real-world tests conducted on the selected traffic routes of a Tier-II Indian city (Dehradun city) using a data logger with multiple sensors. It investigates the effect of real-world driving and different drive-modes (idling, acceleration, deceleration, and cruising) on the EC and performance of the EV.

Abstract Urban road intersections are the most common complex traffic conditions, especially in intersections without signal lights, when a right-of-way conflict occurs between vehicles, the future trajectory of vehicles is full of uncertainty due to the driver’s personalized driving style and the difference in recognition of relevant driving rules. Accurate prediction of vehicle trajectory is of great significance to collision avoidance decision-making and path planning of ADAS and the driverless car. This article proposes a vehicle trajectory prediction method for intersections without signal lights, which combines the traditional vehicle Constant Turn Rate and Acceleration (CTRA) model in the Long Short-Term Memory (LSTM) network by attention mechanism, and the German open-source dataset inD for intersections without signal lights is used to verify the method studied in this article.

Abstract This article introduces a novel approach to maximize the harvested energy during regenerative braking in electric vehicles (EVs) through an optimal distribution between regenerative and friction braking. In the proposed method, the optimum operating point of the electric motor during braking is determined based on the driver’s brake request and real-time vehicle speed. To this end, a maximum-current curve (MCC), which represents the maximum current that can be recaptured during braking for a given resistive torque, is introduced, and the electric motor operating point is maintained on this curve. In this method, energy extraction during the regenerative braking process is maximized due to operation on the MCC, which also leads to an extended effective range for regenerative braking compared to traditional methods. The effectiveness of the proposed approach is experimentally investigated on a hardware-in-the-loop (HIL) EV testbench.

Abstract Sideswipe accidents occur primarily when drivers attempt an improper lane change, drift out of lane, or the vehicle loses lateral traction. In this article, a fusion approach is introduced that utilizes data from two differing modality sensors (a front-view camera and an onboard diagnostics (OBD) sensor) for the purpose of detecting driver’s behavior of lane changing. For lane change detection, both feature-level fusion and decision-level fusion are examined by using a collaborative representation classifier (CRC). Computationally efficient detection features are extracted from distances to the detected lane boundaries and vehicle dynamics signals. In the feature-level fusion, features generated from two differing modality sensors are merged before classification, while in the decision-level fusion, the Dempster-Shafer (D-S) theory is used to combine the classification outcomes from two classifiers, each corresponding to one sensor.

Abstract Parking an articulated vehicle is a challenging task that requires skill, experience, and visibility from the driver. An automatic parking system for articulated vehicles can make this task easier and more efficient. This article proposes a novel method that finds an optimal path and controls the vehicle with an innovative method while considering its kinematics and environmental constraints and attempts to mathematically explain the behavior of a driver who can perform a complex scenario, called the articulated vehicle park maneuver, without falling into the jackknifing phenomena. In other words, the proposed method models how drivers park articulated vehicles in difficult situations, using different sub-scenarios and mathematical models.

Abstract Unstable articulated vehicles pose a serious threat to the occupants driving them as well as the occupants of the vehicles around them. Articulated vehicles typically experience three types of instability: snaking, jack-knifing, and rollover. An articulated vehicle subjected to any of these instabilities can result in major accidents. In this study a Nonlinear Model Predictive Control (NMPC) that applies brake-based torque vectoring on the trailer is developed to improve the articulated vehicle stability. The NMPC formulation includes tire saturation and applies constraints to prevent rollover. The controller output is a left and right brake force allowing the longitudinal velocity change to be incorporated into the model. Simulations were conducted to instigate snaking and jack-knifing and show the NMPC controller result compared to a simple proportional controller.

Abstract Black spot distribution is a major indicator of urban traffic safety. Identifying the spatial distribution of black spots at urban centers is fundamental to understanding the causes, types, and severity of accidents. Many studies have investigated the spatial distribution of black spots worldwide. However, few studies have investigated the causes, types, and severity of accidents associated with population size, land use patterns, and prevailing traffic conditions in black spots. This study attempts to explore these associations in the black spots in the Dammam Metropolitan Area (DMA). The data was collected between 2015 and 2019 from the Traffic Police Department and the Department of Transportation of the Saudi American Oil Company. The data was analyzed using several spatial analytical techniques, including statistical and GIS-based techniques.

Abstract With the increase of electric vehicles on the roads, there is also an increase with vehicles that use regenerative braking (RB). This novel braking method differs from traditional service braking (SB) because RB decelerates the moment the driver releases the accelerator pedal and continues to actively brake if neither pedal is depressed. Since the vehicle actively decelerates when neither pedal is depressed in a vehicle with RB, we hypothesized that this would result in a difference in driver foot behavior. There were two pieces to explore this potential difference. The first piece was to explore time-based measures. The first measure was the time period from when the lead vehicle brake lights illuminate, to when the driver releases the accelerator pedal. The second measure was the time period from when the driver releases the accelerator pedal, to when the driver presses the brake pedal.

Abstract Seatbelts have been acknowledged to be among the most effective vehicle implements that enhance vehicle occupants’ safety. Using seatbelts has been established as a highly effective means of reducing crash severity. On the contrary, speeding has been associated with an increased likelihood of crash occurrence and severity. Investigating factors associated with these two aspects of driving behavior is vital to improving road safety. This study examines the association of previous crash-involvements with seatbelt use and speeding habits by investigating whether crash-involved drivers were less likely to use seatbelts and more likely to adopt speeding habits. The study further explores the effects of annual driving distance on seatbelt use and speeding behaviors, and whether these effects are influenced by previous crash-involvements.

Abstract As pedestrians are among the most critical road users, this research analyzes their vulnerability characteristics and predicts the injury severity of pedestrian crashes through decision tree techniques, rather than using statistical regression models that have particular predefined causal relationships between dependent and independent variables. Five years have been studied in Nablus Governorate/Province (2012–2016), one of 16 governorates in Palestine, as a case study based on reported crash frequencies for developing countries. Tree techniques (CART [Classification and Regression Tree] and CHAID [Chi-Square Automatic Interaction Detector]) were applied to extract the main impacting factors on injury severity for pedestrian crashes. The main contributions considered a small regional context in developing countries and found differences between the results of various methods in injury severity. Fourteen independent variables have been analyzed.

Abstract This research examined tractor operators’ daily vibration exposure A(8) with different input riding parameters, i.e., average speed (m/s) (2.78, 3.89, 5.0), body mass (BM) (kg/m2) (35.3, 32.6, 25.4), and different terrain types (brick, farm, and tar roads). To arrange the systematic sequence of experiments, Taguchi’s L9 orthogonal array has been selected for this study. The signal-to-noise ratio (SNR) is calculated to analyze the overall influence of input parameters over the output parameters. In this study, it is found that A(8) responses exceeded the recommended action value among all the tractor operators according to ISO 2631-1 (1997). The average speeds and various terrain conditions were shown to be the most influential significant variables (p ≤ 0.05), with percentage contributions of 53.71% and 11.53%, respectively.

Abstract The levels of sustained vehicle automation, as recently updated by SAE in J3016 (status: 06/2018), have become common knowledge. They facilitate overall understanding of the issue. Sustained automation describes the shift in workload from purely human-driven vehicles to full automation. Duties of the driver are assigned to the machine as automation levels rise. Yet sustained driving automation does not cover “automated driving” as a whole. Automated driving functions operating on a nonsustained basis cannot be classified by means of levels describing continuous automation. Emergency braking, e.g., is obviously an intensive, but discontinuous, automation of a single task. It cannot be classified under the regime of sustained automation. The resulting lack of visibility of these important functions cannot satisfy - especially in the light of effect they take on traffic safety.

Abstract In order to improve the safety of vehicles driving in mountainous areas and other curve roads, a DCT shift strategy for different drivers is designed in this article. First, the test road is digitized and an electronic map model is built based on particle filtering (PF) to achieve the optimal prediction of road conditions. Then, a driver’s intention recognition model is constructed based on principle component analysis (PCA) and Kohonen neural network (KohonenNN) to accommodate different driving intentions, and the appropriate downshift schedules are designed for different driver’s intentions. In addition, we propose the concept of the curve safety factor, and the current safety level is determined by vehicle speed and distance. By combining the earlier discussed concepts, the predictive curve shift strategy is presented.

Abstract In order to accurately represent the response of live occupants during pre-crash events and frontal crashes, computational human body models (HBMs) that incorporate active musculature must be validated with appropriate volunteer data that represents a wide range of demographic groups and potential crash conditions. The purpose of this study was to quantify and compare occupant kinematic responses for unaware (relaxed) small female and midsize male volunteers during low-speed frontal and frontal-oblique sled tests across multiple test conditions, while recognizing, assessing, and accounting for potential acclimation effects due to multiple exposures. Six 5th percentile female and six 50th percentile male volunteers were exposed to multiple low-speed frontal and frontal-oblique sled tests on two separate test days. Volunteers experienced one test orientation and two pulse severities (1 g and 2.5 g) on each test day.

Abstract An improved control method of automatic emergency braking (AEB) for rear-end collision avoidance is proposed, which combines the advantages of a time-to-collision (TTC) control algorithm and professional driver emergency braking behavior. The TTC control algorithm mostly adopts phased braking, and although it can avoid collision effectively, the braking process is radical and brake comfort is poor. The emergency braking system with professional driver fitting (PDF) has good comfort and can also avoid collision successfully. However, its brake trigger time is too early, which leads to the stopping distance being too large under high-speed conditions and affects the road utilization. By combining the advantages of the two control methods, an improved control algorithm for AEB is proposed. When the TTC value is not greater than a predetermined limit, the PDF control switch will be closed to avoid collision.