Search Results

Due to the high cost of torque sensors, a calculation model of transient torque is required for real-time coordinating control purpose, especially in hybrid electric powertrains. This paper presents a feedforward calculation method based on mean value model of turbocharged non-EGR diesel engines. A fitting variable called fuel coefficient is defined in an affine relation between brake torque and fuel mass. The fitting of fuel coefficient is simplified to depend only on three variables (engine speed, boost pressure, injected fuel mass). And a two-layer feedforward neural network is utilized to fit the experimental data. The model is validated by load response test and ETC (European Transient Cycle) transient test. The RMSE (root mean square error) of the brake torque is less than 3%.

Self-piercing rivet (SPR) joints are a key joining technology for lightweight materials, and they have been widely used in automobile manufacturing. Manual visual crack inspection of SPR joints could be time-consuming and relies on high-level training for engineers to distinguish features subjectively. This paper presents a novel machine learning-based crack detection method for SPR joint button images. Firstly, sub-images are cropped from the button images and preprocessed into three categories (i.e., cracks, edges and smooth regions) as training samples. Then, the Artificial Neural Network (ANN) is chosen as the classification algorithm for sub-images. In the training of ANN, three pattern descriptors are proposed as feature extractors of sub-images, and compared with validation samples. Lastly, a search algorithm is developed to extend the application of the learned model from sub-images into the original button images.

This paper presents a design process with data mining technique and advanced optimization strategy. The proposed design method provides insights in three aspects. First, data mining technique is employed for analysis to identify key factors of design variables. Second, relationship between multiple types of size and shape design variables and performance responses can be analyzed. Last but not least, design preference can be initialized based on data analysis to provide priori guidance for the starting design points of optimization algorithm. An exhaust system design problem which largely contributes to the improvement of vehicular Noise, Vibration and Harshness (NVH) performance is employed for the illustration of the process. Two types of design parameters, structural variable (gauge of component) and layout variable (hanger location), are considered in the studied case.

Designing a vehicle for superior crash safety performance in consumer rating tests such as US-NCAP is a compelling target in the design of passenger vehicles. In today’s context, there is also a high emphasis on making a vehicle as lightweight as possible which calls for an efficient design. In modern vehicle design, these objectives can only be achieved through Computer-Aided Engineering (CAE) for which a detailed CAD (Computer-Aided Design) model of a vehicle is a pre-requisite. In the absence of the latter (i.e. a matured CAD model) at the initial and perhaps the most crucial phase of vehicle body design, a rational approach to design would be to resort to a knowledge-based methodology which can enable crash safety assessment of an assumed design using artificial intelligence techniques such as neural networks.

Driving style and skill classification are of great significance in human-oriented advanced driver-assistance system (ADAS) development. In this paper, we propose Fully-Connected Deep Neural Networks (FC-DNN) to classify drivers’ styles and skills with naturalistic driving data. Followed by the data collection and pre-processing, FC-DNN with a series of deep learning optimization algorithms are applied. In the experimental part, the proposed model is validated and compared with other commonly used supervised learning methods including the k-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT), random forest (RF), and multilayer perceptron (MLP). The results show that the proposed model has a higher Macro F1 score than other methods. In addition, we discussed the effect of different time window sizes on experimental results. The results show that the driving information of 1s can improve the final evaluation score of the model.

The accuracy of tire forces directly affects the vehicle dynamics model precision and determines the ability of the model to develop the simulation platform or design the control strategy. In the high slip angle, due to the complex interactions at tire-road interfaces, the forces generated by the tires are high nonlinearity and uncertainty, which pose issues in calculating tire force accurately. This paper presents a hybrid physical and data-driven tire force calculation framework, which can satisfy the high nonlinearity and uncertainty condition, improve the model accuracy and effectively leverage prior knowledge of physical laws. The parameter identification for the physical tire model and the data-based compensation for the unknown errors between the physical tire model and actual tire force data are contained in this framework. First, the parameters in the selected combined-slip Burckhardt tire model are identified by the nonlinear least square method with tire test data.

An autonomous vehicle is a comprehensive intelligent system that includes environment sensing, vehicle localization, path planning and decision-making control, of which environment sensing technology is a prerequisite for realizing autonomous driving. In the early days, vehicles sensed the surrounding environment through sensors such as cameras, radar, and lidar. With the development of 5G technology and the Vehicle-to-everything (V2X), other information from the roadside can also be received by vehicles. Such as traffic jam ahead, construction road occupation, school area, current traffic density, crowd density, etc. Such information can help the autonomous driving system understand the current driving environment more clearly. Vehicles are no longer limited to areas that can be sensed by sensors. Vehicles with different autonomous driving levels have different adaptability to the environment.

Transport vehicles consume a large amount of fuel with low efficiency, which is significantly affected by drivers' behaviors. An assessment system of eco-driving pattern for buses could identify the deficiencies of driver operation as well as assist transportation enterprises in driver management. This paper proposes an assessment method regarding drivers' economic efficiency, considering driving conditions. To this end, assessment indexes are extracted from driving economy theories and ranked according to their effect on fuel consumption, derived from a database of 135 buses using multiple regression. A layered structure of assessment indexes is developed with application of AHP, and the weight of each index is estimated. The driving pattern score could be calculated with these weights.

Tracked vehicles are widely used for agriculture, construction and many other areas. Due to high emissions, hybrid electric driveline has been applied to tracked vehicles. The hybrid powertrain design for the tracked vehicle has been researched for years. Different from wheeled vehicles, the tracked vehicle not only requires high mobility while straight driving, but also pursues strong steering performance. The paper takes the hybrid track-type dozers (TTDs) as an example and proposes an optimal design of a novel power-split powertrain for TTDs. The commercial hybrid TTD usually adopts the series hybrid powertrain, and sometimes with an extra steering mechanism, which has led to low efficiency and made the structure more complicated. The proposed three-planetary-gear power-split hybrid powertrain can overcome the problems above by utilizing the characteristics of planetary gear sets.

Forecasting the motion of the leading vehicle is a critical task for connected autonomous vehicles as it provides an efficient way to model the leading-following vehicle behavior and analyze the interactions. In this study, a personalized time-series modeling approach for leading vehicle trajectory prediction considering different driving styles is proposed. The method enables a precise, personalized trajectory prediction for leading vehicles with limited inter-vehicle communication signals, such as vehicle speed, acceleration, space headway, and time headway of the front vehicles. Based on the learning nature of human beings that a human always tries to solve problems based on grouping and similar experience, three different driving styles are first recognized based on an unsupervised clustering with a Gaussian Mixture Model (GMM).

The RON(Research Octane Number) is the most important indicator of motor petrol, and the petrol refining process is one of the important links in petrol production. However, RON is often lost during petrol refining and RON Loss means the value of RON lost during petrol refining. The prediction of the RON loss of petrol during the refining process is helpful to the improvement of petrol refining process and the processing of petrol. The traditional RON prediction method relied on physical and chemical properties, and did not fully consider the high nonlinearity and strong coupling relationship of the petrol refining process. There is a lack of data-driven RON loss models. This paper studies the construction of the RON loss model in the petrol refining process.

Velocity prediction on hilly road can be applied to the energy-saving predictive control of intelligent vehicles. However, the existing methods do not deeply analyze the difference and diversity of road slope driving characteristics, which affects prediction performance of some prediction method. To further improve the prediction performance on road slope, and different road slope driving features are fully exploited and integrated with the common prediction method. A rolling prediction-based multi-scale fusion prediction considering road slope transition driving characteristics is proposed in this study. Amounts of driving data in hilly sections were collected by the advanced technology and equipment. The Markov chain model was used to construct the velocity and acceleration joint state transition characteristics under each road slope transition pair, which expresses the obvious driving difference characteristics when the road slope changes.

The objective of the research into modeling and simulation was to provide an improvement to the Wayne State EcoCAR 2 team’s math-based modeling and simulation tools for hybrid electric vehicle powertrain analysis, with a goal of improving the simulation results to be less than 10% error to experimental data. The team used the modeling and simulation tools for evaluating different outcomes based on hybrid powertrain architecture changes (hardware), and controls code development and testing (software). The first step was model validation to experimental data, as the plant models had not yet been validated. This paper includes the results of the team’s work in the U.S. Department of Energy’s EcoCAR 2 Advanced vehicle Technical Competition for university student teams to create and test a plug-in hybrid electric vehicle for reducing petroleum oil consumption, pollutant emissions, and Green House Gas (GHG) emissions.

Vehicle lane keeping system is becoming a new research focus of drive assistant system except adaptive cruise control system. As we all known, vehicle lateral dynamics show strong nonlinear and time-varying with the variety of longitudinal velocity, especially tire’s mechanics characteristic will change from linear characteristic under low speed to strong nonlinear under high speed. For this reason, the traditional PID controller and even self-tuning PID controller, which need to know a precise vehicle lateral dynamics model to adjust the control parameter, are too difficult to get enough accuracy and the ideal control quality. Based on neural network’s ability of self-learning, adaptive and approximate to any nonlinear function, an adaptive PID control algorithm with BP neural network self-tuning online was proposed for vehicle lane keeping.

With the development of intelligent vehicle and active vehicle safety systems, the demand of sensors is increasing, especially in-tire sensors. Tire parameters are essential for vehicle dynamic control, including tire pressure, tire temperature, slip angle, longitudinal force, etc.. The diversification and growth of in-tire sensors require adequate power supply. Traditionally, embedded batteries are used to power sensors in tire, however, they must be replaced periodically because of the limited energy storage. The power limitation of the batteries would reduce the real-time data transmission frequency and deteriorate the vehicle safety. Heightened interest focuses on generating power through energy harvesting systems in replace of the batteries. Current in-tire energy harvesting devices include piezoelectric, electromagnetic, electrostatic and electromechanical mechanism, whose energy sources include tire deformations, vibrations and rotations.

This paper presents experimental investigations of diagnosing and analyzing the low-frequency, low- SNR (Signal to Noise Ratio) noise sources of three motorcycles using a hybrid technology that consists of a passive SODAR (Sonic Detection And Ranging) and modified HELS (Helmholtz Equation Least Squares) methods. The former enables one to determine the precise locations of multiple sound sources in 3D space simultaneously over the entire frequency range that is consistent with a measurement microphone in non-ideal environment, where there are random background noise and unknown interfering signals. The latter enables one to reconstruct all acoustic quantities such as the acoustic pressure, acoustic intensity, time-averaged acoustic power, radiation patterns, and sound transmission paths through arbitrarily shaped vibrating structures.

Intelligent driving, aimed for collision avoidance and self-navigation, is mainly based on environmental sensing via radar, lidar and/or camera. While each of the sensors has its own unique pros and cons, camera is especially good at object detection, recognition and tracking. However, unpredictable environmental illumination can potentially cause misdetection or false detection. To investigate the influence of illumination conditions on detection algorithms, we reproduced various illumination intensities in a photo-realistic virtual world, which leverages recent progress in computer graphics, and verified vehicle detection effect there. In the virtual world, the environmental illumination is controlled precisely from low to high to simulate different illumination conditions in the driving scenarios (with relative luminous intensity from 0.01 to 400). Sedan cars with different colors are modelled in the virtual world and used for detection task.

Clarifying accident causations can provide a strong foundation to prevent traffic accidents and reduce severities. This paper uses Chinese government census data from 1996-2003[1∼8] and models a relationship between various kinds of traffic accident causations and the severities of the traffic accidents based on neural network combined (NNC). The paper adapts multi-folder cross validation concept to enhance the properties of NNC. It then conducts sensitivity analysis on the trained NNC to identify the prioritized importance of traffic accident causations as they are to the severities of traffic accident. Lastly, the results are validated and compared by the findings of previous researches.

In order to improve the drivability and reduce the clutch friction loss, low-cost slope sensor is used in hill-start control of AMT vehicles. After the power spectrum analysis of the original signal and the design of the digital filter, the angle of the slope is obtained with short enough delay and small enough noise. By using this slope angle information, slope resistance force can be calculated online so that the vehicle can be prevented from sliding backward and optimal launch control can be realized. The digital filter of slope angle signal and the optimal controller of dry clutch engagement are embedded in the TCU (Transmission Control Unit) of a micro-car Geely Panda. Real-vehicle experiments are carried out with optimal clutch controller, which shows that the hill-start with low-cost slope sensor and optimal clutch controller can provide successful vehicle launch with little driveline shock. In addition, it can also avoid backward sliding and engine over-speed effectively.

An automotive powertrain system consists of several interactive and linked nonlinear systems. This research focuses on the coordination of Gasoline Direct Injection (GDI) engine, transmission and emission aftertreatment systems. The goal is to design an optimal control strategy for driving performance, emissions (HC, CO, NOX), fuel economy and smoothness when switching engine mode and when shifting gears, under both discrete and continuous limitations. A multivariable control strategy is used to compromise among all powertrain subsystems to achieve optimal overall performance. A nonlinear discrete dynamic programming approach is proposed for hybrid system optimization. The complex multivariable automotive control problem is then simplified into an optimization problem. The feasibility of automotive hybrid control via the discrete dynamic programming approach is demonstrated by results from many numerical simulations under different operating conditions.