Search Results

A Machine Learning-Genetic Algorithm (ML-GA) approach was developed to virtually discover optimum designs using training data generated from multi-dimensional simulations. Machine learning (ML) presents a pathway to transform complex physical processes that occur in a combustion engine into compact informational processes. In the present work, a total of over 2000 sector-mesh computational fluid dynamics (CFD) simulations of a heavy-duty engine were performed. These were run concurrently on a supercomputer to reduce overall turnaround time. The engine being optimized was run on a low-octane (RON70) gasoline fuel under partially premixed compression ignition (PPCI) mode. A total of nine input parameters were varied, and the CFD simulation cases were generated by randomly sampling points from this nine-dimensional input space. These input parameters included fuel injection strategy, injector design, and various in-cylinder flow and thermodynamic conditions at intake valve closure (IVC).

This study pertains to motion control algorithms using statistical calculations based on relative displacement measurements, in particular where the rattle space is strictly limited by fixed end-stops and a load leveling system that allows for roll to go undetected by the sensors. One such application is the cab suspension of semi trucks that use widely-spaced springs and dampers and a load leveling system that is placed between the suspensions, near the center line of the cab. In such systems it is possible for the suspension on the two sides of the vehicle to settle at different ride heights due to uneven loading or the crown of the road. This paper will compare the use of two moving average signals (one positive and one negative) to the use of one root mean square (RMS) signal, all calculated based on the relative displacement measurement.

This paper presents a unified modeling environment to simulate vehicle driving and powertrain operations within the context of the surrounding environment, including interactions between vehicles and between vehicles and the road. The goal of this framework is to facilitate the analysis of the energy impacts of vehicle connectivity and automation, as well as the development of eco-driving algorithms. Connectivity and automation indeed provide the potential to use information about the environment and future driving to minimize energy consumption. To achieve this goal, the designers of eco-driving control strategies need to simulate a wide range of driving situations, including the interactions with other vehicles and the infrastructure in a closed-loop fashion.

As connectivity and sensing technologies become more mature, automated vehicles can predict future driving situations and utilize this information to drive more energy-efficiently than human-driven vehicles. However, future information beyond the limited connectivity and sensing range is difficult to predict and utilize, limiting the energy-saving potential of energy-efficient driving. Thus, we combine a conventional speed optimization planner, developed in our previous work, and reinforcement learning to propose a real-time intelligent speed optimization planner for connected and automated vehicles. We briefly summarize the conventional speed optimization planner with limited information, based on closed-form energy-optimal solutions, and present its multiple parameters that determine reference speed trajectories.

Animal-vehicle collision (AVC) is a significant safety issue on American roads. Each year approximately 1.5 million AVCs occur in the U.S., the majority of them involving deer. The increasing use of cameras and radar on vehicles provides opportunities for prevention or mitigation of AVCs, particularly those involving deer or other large animals. Developers of such AVC avoidance/mitigation systems require information on the behavior of encountered animals, setting characteristics, and driver response in order to design effective countermeasures. As part of a larger study, naturalistic driving data were collected in high AVC incidence areas using 48 participant-owned vehicles equipped with data acquisition systems (DAS). Continuous driving data including forward video, location information, and vehicle kinematics were recorded. The respective 11TB dataset contains 35k trips covering 360K driving miles.

The Unmanned Ground Vehicle Competition is jointly sponsored by the SAE, the Association for Unmanned Vehicle Systems (AUVS), and Oakland University. College teams, composed of both undergraduate and graduate students, build autonomous vehicles that compete by navigating a 139 meter outdoor obstacle course. The course, which includes a sand pit and a ramp, is defined by painted continuous or dashed boundary lines on grass and pavement. The obstacles are arbitrarily placed, multi-colored plastic-wrapped hay bales. The vehicles must be between 0.9 and 2.7 meters long and less than 1.5 meters wide. They must be either electric-motor or combustion-engine driven and must carry a 9 kilogram payload. All computational power, sensing and control equipment must be carried on board the vehicle. The technologies employed are applicable in Intelligent Transportation Systems (ITS).

This research investigates the energy savings achieved through eco-driving controls in connected and automated vehicles (CAVs), with a specific focus on the influence of powertrain characteristics. Eco-driving strategies have emerged as a promising approach to enhance efficiency and reduce environmental impact in CAVs. However, uncertainty remains about how the optimal strategy developed for a specific CAV applies to CAVs with different powertrain technologies, particularly concerning energy aspects. To address this gap, on-track demonstrations were conducted using a Chrysler Pacifica CAV equipped with an internal combustion engine (ICE), advanced sensors, and vehicle-to-infrastructure (V2I) communication systems, compared with another CAV, a previously studied Chevrolet Bolt electric vehicle (EV) equipped with an electric motor and battery.

One of the principal goals of modern vehicle control systems is to ensure passenger safety during dangerous maneuvers. Their effectiveness relies on providing appropriate parameter inputs. Tire-road contact forces are among the most important because they provide helpful information that could be used to mitigate vehicle instabilities. Unfortunately, measuring these forces requires expensive instrumentation and is not suitable for commercial vehicles. Thus, accurately estimating them is a crucial task. In this work, two estimation approaches are compared, an observer method and a neural network learning technique. Both predict the lateral and longitudinal tire-road contact forces. The observer approach takes into account system nonlinearities and estimates the stochastic states by using an extended Kalman filter technique to perform data fusion based on the popular bicycle model.

The availability of active safety systems, such as Lane Departure Warning (LDW), has recently been added as a rating factor in the U.S. New Car Assessment Program (NCAP). The objective of this study is to determine the relevance of the NCAP LDW confirmation test to real-world road departure crashes. This study is based on data collected as part of supplemental crash reconstructions performed on 890 road departure collisions from the National Automotive Sampling System, Crashworthiness Data System (NASS/CDS). Scene diagrams and photographs were examined to determine the lane departure and lane marking characteristics not available in the original data. The results suggest that the LDW confirmation test captures many of the conditions observed in real-world road departures. For example, 40% of all single vehicle collisions in the dataset involved a drift-out-of-lane type of departures represented by the test.

Climate change due to greenhouse gas emissions has led to new vehicle emissions standards which in turn have led to a call for vehicle technologies to meet these standards. Modeling of vehicle fuel consumption and emissions emerged as an effective tool to help in developing and assessing such technologies, to help in predicting aggregate vehicle fuel consumption and emissions, and to complement traffic simulation models. The paper identifies the current state of the art on vehicle fuel consumption and emissions modeling and its utilization to test the environmental impact of the Intelligent Transportation Systems (ITS)’ measures and to evaluate transportation network improvements. The study presents the relevant models to ITS in the key classifications of models in this research area. It demonstrates that the trends of vehicle fuel consumption and emissions provided by current models generally do satisfactorily replicate field data trends.

Reductions in vehicle drive losses are as important to improving fuel economy as increases in powertrain efficiencies. In order to measure vehicle fuel economy, chassis dynamometer testing relies on accurate road load determinations. Road load is currently determined (with some exceptions) using established test track coastdown testing procedures. Because new vehicle technologies and usage cases challenge the accuracy and applicability of these procedures, on-road experiments were conducted using axle torque sensors to address the suitability of the test procedures in determining vehicle road loads in specific cases. Whereas coastdown testing can use vehicle deceleration to determine load, steady-state testing can offer advantages in validating road load coefficients for vehicles with no mechanical neutral gear (such as plug-in hybrid and electric vehicles).

This paper investigates the aerodynamic influence of multiple on-highway trucks in different platooning configurations. Complex pressure fields are generated on the highways due to interference of multiple vehicles. This pressure field causes an aerodynamic drag to be different than the aerodynamic drag of a vehicle in a no-traffic condition. In order to study the effect of platooning, three-dimensional modeling and numerical simulations were performed using STAR-CCM+® commercial Computational Fluid Dynamics (CFD) tool. The aerodynamic characteristics of vehicles were analyzed in five different platooning configurations with two and three vehicles in single and multiple lanes. A significant Yaw Averaged Aerodynamic Drag (YAD) reduction was observed in both leading and trailing vehicles. YAD was based on the average result of three different yaw angles at 0°, −6° and 6°. In single-lane traffic, YAD reduction was up to 8% and 38% in leading and trailing vehicles, respectively.

The combustion system of a heavy-duty diesel engine operated in a gasoline compression ignition mode was optimized using a CFD-based response surface methodology and a machine learning genetic algorithm. One common dataset obtained from a CFD design of experiment campaign was used to construct response surfaces and train machine learning models. 128 designs were included in the campaign and were evaluated across three engine load conditions using the CONVERGE CFD solver. The design variables included piston bowl geometry, injector specifications, and swirl ratio, and the objective variables were fuel consumption, criteria emissions, and mechanical design constraints. In this study, the two approaches were extensively investigated and applied to a common dataset. The response surface-based approach utilized a combination of three modeling techniques to construct response surfaces to enhance the performance of predictions.

Connected and automated technologies poised to change the way vehicles operate are starting to enter the mainstream market. Methods to accurately evaluate these technologies, in particular for their impact on safety and energy, are complex due to the influence of static and environmental factors, such as road environment and traffic scenarios. Therefore, it is important to develop modeling and testing frameworks that can support the development of complex vehicle functionalities in a realistic environment. Microscopic traffic simulations have been increasingly used to assess the performance of connected and automated vehicle technologies in traffic networks. In this paper, we propose and apply an evaluation method based on a combination of microscopic traffic simulation (AIMSUN) and a chassis dynamometer-based vehicle-in-the-loop environment, developed at Argonne National Laboratory.

This paper presents the energy savings of an automated driving control applied to an electric vehicle based on the on-track testing results. The control is a universal speed planner that analytically solves the eco-driving optimal control problem, within a receding horizon framework and coupled with trajectory tracking lower-level controls. The automated eco-driving control can take advantage of signal phase and timing (SPaT) provided by approaching traffic lights via vehicle-to-infrastructure (V2I) communications. At each time step, the controller calculates the accelerator and brake pedal position (APP/BPP) based on the current state of the vehicle and the current and future information about the surrounding environment (e.g., speed limits, traffic light phase).

As emerging automated vehicle technology is making advances in safety and reliability, engineers are also exploring improvements in energy efficiency with this new paradigm. Powertrain efficiency receives due attention, but also impactful is finding ways to reduce driving losses in coordinated-driving scenarios. Efforts focused on simulation to quantify road load improvements require a sufficient amount of background validation work to support them. This study uses a practical approach to directly quantify road load changes by testing the coordinated driving of two vehicles on a test track at various speeds (64, 88, 113 km/h) and vehicle time gaps (0.3 to 1.3 s). Axle torque sensors were used to directly measure the load required to maintain steady-state speeds while following a lead vehicle at various gap distances.

Structural Health Monitoring (SHM), especially in the field of rotary machinery diagnosis, plays a crucial role in determining the defect category as well as its intensity in a machine element. This paper proposes a new framework for real-time classification of structural defects in a roller bearing test rig using time domain-based classification algorithms. Along with the bearing defects, the effect of eccentric shaft loading has also been analyzed. The entire system comprises of three modules: sensor module – using accelerometers for data collection, data processing module – using time-domain based signal processing algorithms for feature extraction, and classification module – comprising of deep learning algorithms for classifying between different structural defects occurring within the inner and outer race of the bearing.

Over 250 million vehicles are operating on United States roads and highways and over 12 million of them reach the end of their useful lives annually. These end-of-life vehicles (ELVs) contain over 24 million tons (21.8 million metric tonnes) of materials including ferrous and non-ferrous metals, polymers, glass, and automotive fluids. They also contain many parts and components that are still useable and some that could be economically rebuilt or remanufactured. Dismantlers acquire the ELVs and recover from them parts for resale “as-is” or after remanufacturing. The dismantler then sells what remains of the vehicle, the “hulk”, to a shredder who shreds it to recover and sell the metals. Presently, the remaining non-metallic materials, commonly known as shredder residue, are mostly landfilled. The vehicle manufacturers, now more than ever, are working hard to build more energy efficient and safer, more affordable vehicles.

The road profile has been shown to have significant effects on various vehicle conditions including ride, handling, fatigue or even energy efficiency; as a result it has become a variable of interest in the design and control of numerous vehicle parts. In this study, an integrated state estimation algorithm is proposed that can provide continuous information on road elevation and profile variations, primarily to be used in active suspension controls. A novel tire instrumentation technology (smart tire) is adopted together with a sensor couple of wheel attached accelerometer and suspension deflection sensor as observer inputs. The algorithm utilizes an adaptive Kalman filter (AKF) structure that provides the sprung and unsprung mass displacements to a sliding-mode differentiator, which then yields to the estimation of road elevations and the corresponding road profile along with the quarter car states.

Modern vehicles use automated driving assistance systems (ADAS) products to automate certain aspects of driving, which improves operational safety. In the U.S. in 2020, 38,824 fatalities occurred due to automotive accidents, and typically about 25% of these are associated with inclement weather. ADAS features have been shown to reduce potential collisions by up to 21%, thus reducing overall accidents. But ADAS typically utilize camera sensors that rely on lane visibility and the absence of obstructions in order to function, rendering them ineffective in inclement weather. To address this research gap, we propose a new technique to estimate snow coverage so that existing and new ADAS features can be used during inclement weather. In this study, we use a single camera sensor and historical weather data to estimate snow coverage on the road. Camera data was collected over 6 miles of arterial roadways in Kalamazoo, MI.