Timothy Gordon

Stochastic Modeling for Studies of Real-World PHEV Usage: Driving Schedule and Daily Temporal Distributions

Daily driving missions provide the fundamental information required to predict the impact of the plug-in hybrid electric vehicle (PHEV) on the grid. In this paper, we propose a statistical modeling approach of daily driving mission sets. The approach consists of temporal distribution modeling and the synthesis of individual representative cycles. The proposed temporal distribution model can capture departure and arrival time distributions with a small number of samples by statistically relating the distributions. Then, representative naturalistic cycles are constructed through a stochastic process and a subsequent statistical analysis with respect to driving distance. They are randomly assigned to the temporal distribution model to build up complete daily driving missions. The proposed approach enables the assessment of the impact on the grid of a large-scale deployment of PHEVs using a small number of simulations capturing real-world driving patterns and the temporal distributions of departure and arrival times.

Characterizing naturalistic driving patterns for Plug-in Hybrid Electric Vehicle analysis

While much of the previous research relies on Federal Driving Schedules originally developed for emission certification tests of conventional vehicles, consumer acceptance and market penetration will depend on PHEV performance under realistic driving conditions. Therefore, characterizing the actual driving is essential for PHEV design and control studies, and for establishing realistic forecasts pertaining to vehicle energy consumption and charging requirements. To achieve this goal, we analyze naturalistic driving data generated in Field Operational Tests (FOT) of passenger vehicles in Southeast Michigan. The FOT were originally conceived for evaluating driver interaction with advanced safety systems, but the databases are rich with information pertaining to vehicle energy. After the initial statistical analysis of the vehicle speed histories, the naturalistic driving schedules are used as input to the PHEV computer simulation to predict energy usage as a function of trip length. The highest specific energy, i.e. energy per mile, is critical for battery and motor sizing. As an illustration of the impact of actual driving, the low-energy and high-energy driving patterns would require PHEV20 battery sizes of 6.12 kWh and 13.6 kWh, respectively. This is determined assuming that the minimum state of charge (SOC) is 40%. In addition, the naturalistic driving databases are mined for information about vehicle resting time, i.e. time spent at typical locations during the 24-hour period. The locations include “home”, “work”, “large-business” such as a large retail store, and “small business”, such as a gas station, and finally “residential” other than home. The characterization of vehicle daily missions supports analysis of charging schedules, as it indicates times spent at given locations as well as the likely battery SOC at the time of arrival.

Automated driving and autonomous functions on road vehicles

In recent years, road vehicle automation has become an important and popular topic for research and development in both academic and industrial spheres. New developments have received extensive coverage in the popular press, and it may be said that the topic has captured the public imagination. Indeed, the topic has generated interest across a wide range of academic, industry and governmental communities, well beyond vehicle engineering; these include computer science, transportation, urban planning, legal, social science and psychology. While this follows a similar surge of interest – and subsequent hiatus – of Automated Highway Systems in the 1990s, the current level of interest is substantially greater, and current expectations are high. It is common to frame the new technologies under the banner of ‘self-driving cars’ – robotic systems potentially taking over the entire role of the human driver, a capability that does not fully exist at present. However, this single vision leads one to ignore the existing range of automated systems that are both feasible and useful. Recent developments are underpinned by substantial and long-term trends in ‘computerisation’ of the automobile, with developments in sensors, actuators and control technologies to spur the new developments in both industry and academia. In this paper, we review the evolution of the intelligent vehicle and the supporting technologies with a focus on the progress and key challenges for vehicle system dynamics. A number of relevant themes around driving automation are explored in this article, with special focus on those most relevant to the underlying vehicle system dynamics. One conclusion is that increased precision is needed in sensing and controlling vehicle motions, a trend that can mimic that of the aerospace industry, and similarly benefit from increased use of redundant by-wire actuators.

Model-based predictive control of vehicle dynamics

This paper proposes a new model-based hierarchical control strategy for the application of brake torque distribution to vehicle stability control. A three-layer control architecture is adopted, and the focus of the paper is on the intermediate layer which employs a form of model predictive control. To achieve this a non-linear vehicle model is linearised at successive time instants about non-equilibrium operating points. Simulations show that the new controller is more efficient (uses reduced brake authority) compared to a conventional rule-based control algorithm. It also has the advantage of being highly flexible for future application in fully integrated chassis systems.

A flexible hierarchical model-based control methodology for vehicle active safety systems

A hierarchical control scheme is applied to the problem of integrated chassis control of a collision avoidance system (CAS). This includes both lateral and longitudinal control, using Active Front Steer in addition to the brake actuators. The inherent flexibility of the control system is provided by the intermediate layer, which employs a form of model predictive control to determine actuator apportionment. The desired vehicle motions in the upper layer, in the form of reference yaw rate and two-dimensional mass center accelerations, are determined using a kinematic policy (KP) for collision avoidance. The KP uses simple information about range and azimuth angles for multiple points that bound the available vehicle trajectory, and prioritises yaw motion response based on the worst case collision threat. This KP approach for CAS is more practical than trajectory tracking approaches because the KP does not need a pre-defined a reference path and does not need any computationally intensive optimisation of the vehicle motion control.

Quasi-Linear Optimal Path Controller Applied to Post Impact Vehicle Dynamics

This paper investigates brake-based path control of a passenger vehicle, aimed at reducing secondary collision risk, following an initial impact in a traffic accident. This risk may be reduced if lateral deviations from the preimpact path can be minimized, at least on straight roads. Numerical optimization has previously shown that coupled control of lateral forces and yaw moments can be applied to effectively minimize such path deviations. In this paper, a quasi-linear optimal controller (QLOC) is proposed to achieve this control target. QLOC uses nonlinear optimal control theory to provide a semiexplicit approximation for optimal post impact (PI) path control. The controller design method is novel, combining linear costate dynamics with nonlinear constraints due to tire friction limits. A fully closed-loop form of the controller is presented; it is applicable to multiple-event accidents occurring on straight roads, including adaptive estimation of the time instant at maximum deviation. The controller achieves performance that is very similar to that of open-loop numerical optimization. Assuming that the vehicle remains on the road surface after the impact and that the brake actuators remain operational, it is verified that the path controller is effective over a wide range of PI kinematic conditions. It is expected that the QLOC controller will prove useful in other cases where chassis systems directly control the vehicle path, e.g., in crash-imminent avoidance maneuvers.

Overview of a cyber-enabled wireless monitoring system for the protection and management of critical infrastructure systems

The long-term deterioration of large-scale infrastructure systems is a critical national problem that if left unchecked, could lead to catastrophes similar in magnitude to the collapse of the I-35W Bridge. Fortunately, the past decade has witnessed the emergence of a variety of sensing technologies from many engineering disciplines including from the civil, mechanical and electrical engineering fields. This paper provides a detailed overview of an emerging set of sensor technologies that can be effectively used for health management of large-scale infrastructure systems. In particular, the novel sensing technologies are integrated to offer a comprehensive monitoring system that fundamentally addresses the limitations associated with current monitoring systems (for example, indirect damage sensing, cost, data inundation and lack of decision making tools). Self-sensing materials are proposed for distributed, direct sensing of specific damage events common to civil structures such as cracking and corrosion. Data from self-sensing materials, as well as from more traditional sensors, are collected using ultra low-power wireless sensors powered by a variety of power harvesting devices fabricated using microelectromechanical systems (MEMS). Data collected by the wireless sensors is then seamlessly streamed across the internet and integrated with a database upon which finite element models can be autonomously updated. Life-cycle and damage detection analyses using sensor and processed data are streamed into a decision toolbox which will aid infrastructure owners in their decision making.

On optimal recovery from terminal understeer

This paper addresses the problem of terminal understeer and its mitigation via integrated brake control. The scenario considered is when a vehicle enters a curve at a speed that is too high for the tyre–road friction limits and an optimal combination of braking and cornering forces is required to slow the vehicle down and to negotiate the curve. Here, the driver commands a step steering input, from which a circular arc reference path is inferred. An optimal control problem is formulated with an objective to minimize the maximum off-tracking from the reference path, and two optimal control solutions are obtained. The first is an explicit analytical solution for a friction-limited particle; the second is a numerically derived open-loop brake control sequence for a nonlinear vehicle model. The particle solution is found to be a classical parabolic trajectory associated with a constant acceleration vector of the global mass center. The independent numerical optimization for the vehicle model is found to approximate closely the kinematics of the parabolic path reference strategy obtained for the particle. Using the parabolic path reference strategy, a closed-loop controller is formulated and verified against the solution from numerical optimization. The results are further compared with understeer mitigation by yaw control, and the parabolic path reference controller is found to give significant improvement over yaw control for this scenario.

Characterizing One-day Missions of PHEVs Based on Representative Synthetic Driving Cycles

This paper investigates series plug-in hybrid electric vehicle (PHEV) behavior during one-day with synthesized representative one-day missions. The amounts of electric energy and fuel consumption are predicted to assess the PHEV impact on the grid with respect to the driving distance and different charging scenarios: (1) charging overnight, (2) charging whenever possible. The representative cycles are synthesized using the extracted information from the real-world driving data in Southeast Michigan gathered through the Field Operational Tests (FOT) conducted by the University of Michigan Transportation Research Institute (UMTRI). The real-world driving data include 4,409 trips covering 830 independent days and temporal distributions of departure and arrival times. The sample size is large enough to represent real-world driving. The driving cycle synthesis approach proposed by Lee, and Filipi 2,3 based on a stochastic process and subsequent validation procedure is used to create real-world driving cycles. To cover the entire range of real-world driving distance, ten synthetic cycles are created ranging from 4.78 miles to 40.71 miles following the real-world driving distance distribution. The PHEV behavior over one-day is characterized through a simulation based approach. The PHEV simulation is executed using Matlab simulink based Powertrain System Analysis Toolkit (PSAT) developed by Argonne National Laboratory (ANL) and in-house Matlab codes. The amounts of the electricity and fuel consumptions over one-day are predicted under different driving distances and different charging scenarios. The prediction of the PHEV behavior can be directly linked to the loads on the local distribution network. © 2011 SAE International.

Modeling human lane keeping control in highway driving with validation by naturalistic data

This paper considers two models of steering control relevant to lane-keeping during normal driving. The first is well-known from the literature and uses linear feedback. To test whether the model is capable of representing real-world driving under conditions of lane-keeping with low workload, parameter fitting is carried out using naturalistic driving data (NDD). It is found that the model can fit the NDD quite well, and that two of the three control parameters may be estimated in a consistent and repeatable way. The lack of fit in the third parameter is explained by the recognition that lane-center tracking does not occur in practice; hence response to lane boundaries is considered more relevant. However the instability of the resulting closed-loop controller calls into question the validity of the linear model. Further, it is found that the linear model does not adequately represent the intermittent pulse-like qualities of real-world steering control. Based on these considerations, a second model is formulated and initial comparisons with NDD are presented. It is proposed that the new model may inherently take account of workload demands, and is therefore relevant to issues of visual attention allocation during driving under reduced workload.