Examining the effect of speed,roadside features,and roadway geometry on crash experience along a rural corridor

This paper presents a current investigation into crash experience along a 15.7-mile rural corridor in southwest Montana with the aim of better understanding crash causal factors along the corridor. The study utilized ten years of crash data, geometric data, and observed freeflow speed data along the corridor. A systematic approach was used where every tenth of a mile was described in term of the crash experience, speed, alignment, and roadside features. Using bivariate and multivariate statistical anal-yses, the study investigated the crash experience along the corridor as well as some of the underlying relationships which could explain some of the crash causal factors. Results show a strong association between crash rates and horizontal curvatures even for flat curves that can be negotiated at speeds above the posted speed limit, per the highway design equations. Higher crash rates were also found to be associated with the difference between the observed free-flow speeds and the speed dictated by the curve radius or sight distance as per the design equations. Further, results strongly support the safety benefits of guardrails as evidenced by the lower crash rates and severities. The presence of fixed objects and the steepness of side slopes were also found to have an effect on crash rates and severities.     

High speed driving stability of passenger car under crosswind effects

The driving stability of a passenger car at high-speed and under crosswind conditions is affected by aerodynamic characteristics as well as their dynamic characteristics, suspension, and weight distribution. In this study, the total measuring system was thought up to understand the transient vehicle dynamics and aerodynamics with driver’s control inputs all together. The test results were taken from a full-scale wind tunnel test, a crosswind generator test and an on-road test. We investigated major aerodynamic parameters that affect the driving stability of passenger cars under crosswind effects such as overtaking, passing each other, natural crosswind, etc. The reaction rate of high-speed stability will be improved when we minimize the total lift, side force and especially the yawing moment.     

Parallel model based fault detection algorithm for electronic parking brake system

This paper describes a parallel model-based fault detection algorithm for an electronic parking brake (EPB) system, which consists of an electronic control unit with built-in current sensor and braking force sensor. For the EPB system to supply sufficient parking force to a vehicle, the parking force sensor is of utmost importance. If a fault occurs in this sensor, sufficient parking force may not be supplied, thereby seriously threatening the safety of the vehicle. Thus, a fault detection method is required for the parking force sensor of the EPB system to improve the safety of vehicles. For this purpose, a highly reliable fault detection method is needed to detect abnormal fault signals, which cannot be detected by the existing on-line sensor monitoring fault detection methods. This paper proposes a novel parallel model-based fault detection algorithm for the EPB system, which compares the physical sensor data with the mathematical model, the fuzzy model, and the neural network model at the same time. In order to reduce false alarms, the magnitude of thresholds and the operation counts are changed adaptively. When the proposed parallel model-based fault detection algorithm detects severe failures of the force sensor, it warns the driver in advance to prevent accidents due to the failures. The proposed algorithm is verified by hardware-in-theloop simulations in various situations.     

In-situ measurement of loading stresses with X-ray diffraction for yield locus determination

The application of the X-ray diffraction method is introduced to solve the problem of inhomogeneous deformation fields in the specimens used for sheet metal characterization. In this method, strains are measured on one side of a specimen with optical measurement systems. On the other side, loading stresses on a specimen are captured with an X-ray diffractometer mounted on a universal testing machine. By this way, the whole stress-strain history of a material point is tracked during testing. The method was first applied to uniaxial tension tests, whereby the applicability of the theory of stress factors and effective X-ray elastic constants were tested. The relaxation behavior of a sheet material which shows itself as stress drops during in-situ experimentation was characterized and compensated by a visco-plastic material model for different stress states. The proposed method was applied to characterize aluminum alloy AA5182 under plane strain tension and shear conditions and the results were compared with the conventionally obtained yield locus. Numerical analyses of a workpiece with the Vegter and Yld2000-2D material models show that the enriched yield locus definition with accurate plane strain tension and shear stresses captures the experimentally obtained surface strains more precisely.     

Innovative design optimization strategy for the automotive industry

In order to effectively solve modern automotive design problems including the results of nonlinear FEA and multi-body dynamics, a progressive meta-model based design optimization is presented. To reduce the number of initial sample points, two sampling methods are introduced. Then, for efficient and stable construction of meta-models, three metamodel methods are newly introduced which are numerically based on the singular value decomposition technique. To design a practical system considering manufacturing tolerances and optimizing multiple performances, a robust design optimization, 6-sigma constraints and multi-objective strategies are implemented when solving the approximate optimization problem constructed from the meta-models. Until the convergence criteria are satisfied, the initially developed meta-models are progressively improved by adding only one point that minimizes the approximate Lagrangian in the consecutive optimization iterations. Finally, one validation sample and four automotive applications are solved to show the effectiveness of the proposed approach.     

Co-operative control for regenerative braking and friction braking to increase energy recovery without wheel lock

This paper presents a regenerative braking co-operative control algorithm to increase energy recovery without wheel lock. Considering the magnitude of the braking force available between the tire and road surface, the control algorithm was designed for the regenerative braking force at the front wheel and friction braking force at the rear wheel to be increased following the friction coefficient line. The performance of the proposed regenerative braking co-operative control algorithm was evaluated by the hardware in the loop simulation (HILS) with an electronic wedge brake on its front wheels and an electronic mechanical brake on its rear wheels. The HILS results showed that a proper braking force on the front and rear wheels on a low μ road prevented the lock of the front wheels that was connected to the motor, and maintained the regenerative braking and increased energy recovery.     

Bumper optimum design using the dynamically equivalent beam under various impact conditions

Nowadays it is required for the bumper system to meet the various impact conditions simultaneously; barrier impact, IIHS (Insurance Institute for Highway Safety) bumper impact and pedestrian impact. Firstly, dynamically equivalent bumper beam models were developed for each impact condition and its accuracy was verified by nonlinear finite element analysis result. Dynamically equivalent pedestrian impact beam model was developed by using the equivalent forces of bumper beam and stiffeners. Pedestrian bending angle was obtained by using this equivalent pedestrian beam model. By combining these equivalent beam models, bumper optimum design program was developed. In this optimum design program, direct search method was used for the optimization algorithm. To verify the accuracy of this optimum design program, a nonlinear finite element result was used. By using this optimum design program, it can be secured the bumper impact performances in an early design stage and it will be also contributed to reduce the design time and test costs.     

Links between subjective assessments and objective metrics for steering

The characteristics of steering perception are decisive factors for overall driver preference and for vehicle safety. Car manufacturers are continuously required to tune the characteristics of the vehicle and have a strong need to be more effective in the design and evaluation of cars. Using only objective metrics (OM) can result in unwanted steering feel and using only subjective assessments (SA) is time-consuming, costly and non-repetitive. Before a tool can be built to predict the steering feel in front-end development and to improve design knowledge from the full vehicle level to the component level, links between subjective assessments and objective metrics must be found and analysed. The data collected for the study presented in this paper include subjective ratings from expert drivers and objective measurements made with steering robots, involving twelve expert drivers and over twenty vehicles across four different vehicle classes. Linear regression and neural network analysis (NN) have been used to explore reliable subjective-objective links. The tools and methods used in this research showed promising results. Most of the links found concern response and torque feedback. The preferred ranges of some crucial objective metrics leading to more desirable steering feel have been defined and presented. The results indicate that it would be possible for car manufacturers to develop new vehicles more effectively with a steering feel in line with the design criteria by using the tools and methods investigated in this paper.     

Minimizing vehicle post impact path lateral deviation using optimized braking and steering sequences

This paper investigates the optimal control of a vehicle, after a light impact during a traffic accident. To reduce the risk of secondary events, the control target is set: to minimize the maximum lateral deviation from the initial path. In previous analysis path control was achieved by the active control of individual wheel braking. The present paper examines potential benefits from the additional control of front steering angles. Numerical optimization is used to determine optimal control sequences for both actuator configurations. It is found that steering provides significant control benefits, though not for all post-impact kinematics. For all cases considered, the optimal control operates at the boundary of the control domain of available forces and moments. This domain is expanded when steering is available, and there exists an expanded range of conditions for which coupled control of yaw moments and lateral forces is the most effective control strategy. The sensitivity of vehicle response to the individual actuator controls is studied; it reveals this sensitivity is related to the actuator bandwidth and the lack of any dynamic cost in the longitudinal direction. This motivates a further analysis which includes longitudinal and lateral dynamics in the cost function. This is broadly related to real-world crash risks. Further, different versions of such cost functions are compared as a basis for implementation in a closed-loop controller.     

New tyre-road contact model for applications at low speed

Most of the tyre models have been developed for high speed, combined forces, etc., however, in certain tests it is necessary to know tyre behaviour at very low speed in order to evaluate different systems. So, during vehicle inspection and maintenance of the steering and brake system, by means of sideslip tester and roller brake tester respectively, the forces transmitted by the tyres are measured; all of these inspections are carried out at low speeds. Furthermore, usually, automobile vehicles run at low speeds during an important part of their operating life (less than 60 km/h), mainly during urban traffic, and in steady state conditions. Therefore, it is particularly interesting to develop an accurate model of the contact patch tyrepavement for low speeds without the complexity of models that cover a wide speed range but provide less precision at very low speeds. The dynamometer plate has proved to be an appropriate test equipment to characterise the tyre-pavement contact at low speed and the steering geometry and wheel alignment. It has the feature of being able to carry out tests with the tyre installed in the vehicle as in completely real conditions. The main aim of this research is to set up a contact model between tyre and pavement at very low speed based on the measurement of longitudinal and lateral forces. A test methodology that allows carrying out the experimental tests in a systematic and controlled way with the dynamometer plate has also been developed. From this model it will be possible to estimate the forces that tyres are capable of transmitting in different situations to act in the parameters which affect these forces and maximize them.