Search Results

The focus of this paper is to discuss the modeling and control of intake plenum pressure on the Powertrain Control Research Laboratory's (PCRL) Single-Cylinder Engine (SCE) transient test system using a patented device known as the Intake Air Simulator (IAS), which dynamically controls the intake plenum pressure, and, subsequently, the instantaneous airflow into the cylinder. The IAS exists as just one of many devices that the PCRL uses to control the dynamic boundary conditions of its SCE transient test system to make it “think” and operate as though it were part of a Multi-Cylinder Engine (MCE) test system. The model described in this paper will be used to design a second generation of this device that utilizes both continuously and discretely actuating valves working in parallel.

There is significant potential for connected and autonomous vehicles to impact vehicle efficiency, fuel economy, and emissions, especially for hybrid-electric vehicles. These improvements could have large-scale impact on oil consumption and air-quality if deployed in large Mobility-as-a-Service or ride-sharing fleets. As part of the US Department of Energy's current Advanced Vehicle Technology Competition (AVCT), EcoCAR: The Mobility Challenge, Mississippi State University’s EcoCAR Team is redesigning and doing the development work necessary to convert a conventional gasoline spark-ignited 2019 Chevy Blazer into a hybrid-electric vehicle with SAE Level 2 autonomy. The target consumer segments for this effort are the Mobility-as-a-Service fleet owners, operators and riders. To accomplish this conversion, the MSU team is implementing a P4 mild hybridization strategy that is expected to result in a 30% increase in fuel economy over the stock Blazer.

Powertrain simulation is becoming an increasingly valuable tool to evaluate new technologies proposed for future military vehicles. The powertrain of the M1A1 Abrams tank is currently being modeled in the Powertrain Control Research Laboratory (PCRL) at the University of Wisconsin-Madison. This powertrain model is to be integrated with other component models in an effort to produce a high fidelity simulation of the entire vehicle.

Neural networks are useful tools for optimization studies since they are very fast, so that while capturing the accuracy of multi-dimensional CFD calculations or experimental data, they can be run numerous times as required by many optimization techniques. This paper describes how a set of neural networks trained on a multi-dimensional CFD code to predict pressure, temperature, heat flux, torque and emissions, have been used by a genetic algorithm in combination with a hill-climbing type algorithm to optimize operating parameters of a diesel engine over the entire speed-torque map of the engine. The optimized parameters are mass of fuel injected per cycle, shape of the injection profile for dual split injection, start of injection, EGR level and boost pressure. These have been optimized for minimum emissions. Another set of neural networks have been trained to predict the optimized parameters, based on the speed-torque point of the engine.

Neural networks have been used for engine computations in the recent past. One reason for using neural networks is to capture the accuracy of multi-dimensional CFD calculations or experimental data while saving computational time, so that system simulations can be performed within a reasonable time frame. This paper describes three methods to improve upon neural network predictions. Improvement is demonstrated for in-cylinder pressure predictions in particular. The first method incorporates a physical combustion model within the transfer function of the neural network, so that the network predictions incorporate physical relationships as well as mathematical models to fit the data. The second method shows how partitioning the data into different regimes based on different physical processes, and training different networks for different regimes, improves the accuracy of predictions.

A procedure has been developed to build system level predictive models that incorporate physical laws as well as information derived from experimental data. In particular a soot model was developed, trained and tested using experimental data. It was seen that the model could fit available experimental data given sufficient training time. Future accuracy on data points not encountered during training was estimated and seen to be good. The approach relies on the physical phenomena predicted by an existing system level phenomenological soot model coupled with ‘weights’ which use experimental data to adjust the predicted physical sub-model parameters to fit the data. This approach has developed from attempts at incorporating physical phenomena into neural networks for predicting emissions. Model training uses neural network training concepts.

Identifying edge cases for designed algorithms is critical for functional safety in autonomous driving deployment. In order to find the feasible boundary of designed algorithms, simulations are heavily used. However, simulations for autonomous driving validation are expensive due to the requirement of visual rendering, physical simulation, and AI agents. In this case, common sampling techniques, such as Monte Carlo Sampling, become computationally expensive due to their sample inefficiency. To improve sample efficiency and minimize the number of simulations, we propose a tailored active learning approach combining the Support Vector Machine (SVM) and the Gaussian Process Regressor (GPR). The SVM learns the feasible boundary iteratively with a new sampling point via active learning. Active Learning is achieved by using the information of the decision boundary of the current SVM and the uncertainty metric calculated by the GPR.

The SAE AutoDrive Challenge II is a four-year collegiate competition dedicated to developing a Level 4 autonomous vehicle by 2025. In January 2023, the participating teams each received a Chevy Bolt EUV. Within a span of five months, the second phase of the competition took place in Ann Arbor, MI. The authors of this contribution, who participated in this event as team Wisconsin Autonomous representing the University of Wisconsin–Madison, secured second place in static events and third place in dynamic events. This has been accomplished by reducing reliance on the actual vehicle platform and instead leveraging physical analogs and simulation. This paper outlines the software and hardware infrastructure of the competing vehicle, touching on issues pertaining sensors, hardware, and the software architecture employed on the autonomous vehicle. We discuss the LiDAR-camera fusion approach for object detection and the three-tier route planning and following systems.