Friday, November 8, 2013

Raw Engineering moved to Twitter

Blogging papers didn't work out, so you can get the latest U-M Engineering research on Twitter instead @UMRawEng.

Friday, August 3, 2012

Self-confirming price-prediction strategies for simultaneous one-shot auctions

Abstract: Bidding in simultaneous auctions is challenging because an agent’s value for a good in one auction may depend on the outcome of other auctions; that is, bidders typically face an exposure problem. Given the gap in understanding (e.g., lack of game-theoretic solutions) of general simultaneous auction games, previous works have tackled the problem of how to bid in these games with heuristic strategies that employ probabilistic price predictions—so-called price-prediction strategies. We introduce a concept of self-confirming prices, and show that within an independent private value model, bidding optimally with respect to self-confirming price predictions is w.l.o.g. in equilibrium. In other words, Bayes-Nash equilibrium can be fully characterized as a profile of optimal price-prediction strategies with self-confirming predictions. We exhibit practical procedures to compute approximately optimal bids given a probabilistic price predicti! on, and near self-confirming price predictions given a price-prediction strategy. We call the output of our procedures self-confirming price-prediction (SCPP) strategies. An extensive empirical game-theoretic analysis demonstrates that SCPP strategies are effective in simultaneous auction games with both complementary and substitutable preference structures.

Impact: This work (a collaboration between Michigan and Brown) addresses a fundamental issue in automated trading: how to deal with multiple related markets at once. Our solution is justified by game-theoretic analysis, yet is computationally practical given a model of the bidding environment.

Conference: Twenty-Eighth Conference on Uncertainty in Artificial Intelligence, August 15-17, Catalina Island, USA.

Paper URL: http://web.eecs.umich.edu/srg/?page_id=1346

Submitted by Michael Wellman, Professor of Electrical Engineering and Computer Science. wellman@umich.edu

Thursday, August 2, 2012

Ethical education of engineering students

Paper Titles: "An Exploratory Investigation of the Ethical Behavior of Engineering Undergraduates," "Framing faculty and student discrepancies in engineering ethics education delivery" and "An assessment of engineering students’ curricular and co-curricular experiences and their ethical development"

Summary: Cindy Finelli, director of the Center for Research on Learning and Teaching in Engineering at U-M, and colleagues Donald Carpenter of Lawrence Technological University, Trevor Harding of California Polytechnic State University and Janel Sutkus of Carnegie Mellon University, have published two articles on research from their 2007 NSF grant on the curricular, co-curricular and environmental determinants of ethical development among engineering undergraduates. The articles were based upon data collected through a 4,000-student national survey and during visits to 18 engineering schools across the country. The team, known as the E3 Team (Exploring Ethical Decision Making in Engineering), recently received a second NSF grant to create a smaller version of the survey, which will be made available to any engineering school that wishes to study their own students’ level of ethical development by evaluating the effects of individual ethics interventions.

Impact: Measuring the incorporation of ethics into the education of engineering students.

Journal: Journal of Engineering Education. Published in print. April and July, 2012

Paper URLs: http://www.jee.org/2012/April/08, http://www.jee.org/2012/April/02http://www.jee.org/2012/July/04

Submitted by Cindy Finelli, Director of the Center for Research on Learning and Teaching in Engineering and Research Associate Professor. cfinelli@umich.edu

Monday, July 2, 2012

Velocity occupancy space: autonomous navigation in an uncertain, dynamic environment

Abstract: In order to autonomously navigate in an unknown environment, a robotic vehicle must be able to sense obstacles, determine their velocities, and select a collision-free path that will lead quickly to a goal. However, the perceived location and motion of the obstacles will be uncertain due to the limited accuracy of the robot’s sensors. Thus, it is necessary to develop a system that can avoid moving obstacles using uncertain sensor data. The method proposed here is based on an occupancy grid – which has been used to avoid stationary obstacles in an uncertain environment – in conjunction with velocity obstacles – which allow a robot to avoid well-known moving obstacles. The combination of these techniques leads to Velocity Occupancy Space (VOS): a search space which allows the robot to avoid moving obstacles and navigate efficiently to a goal using uncertain sensor data. The proposed method is validated by numerous simulation trials.

Impact: Collision-avoidance for unmanned ground vehicles, with low-cost sensors, is important if these vehicles are to be used in close-proximity with humans and human-operated vehicles.

Journal: Published online as an accepted paper in the International Journal of Vehicle Autonomous Systems, anticipated print date 07/31/2012.

Submitted by A. Galip Ulsoy, the C D Mote, Jr Distinguished University Professor of Mechanical Engineering and Professor of Mechanical Engineering, College of Engineering. ulsoy@umich.edu

Tuesday, June 19, 2012

Two-stage multi-scale search for sparse targets

Abstract: We consider the problem of energy constrained and noise-limited search for targets that are sparsely distributed over a large area. We propose a multi-scale search algorithm that significantly reduces the search time of the adaptive resource allocation policy (ARAP) introduced in [Bashan et all, 2008]. Similarly to ARAP, the proposed approach scans a Q-cell partition of the search area in two stages: first the entire domain is scanned and second a subset of the domain, suspected of containing targets, is re-scanned. The search strategy of the proposed algorithm is driven by maximization of a modified version of the previously introduced ARAP objective function, which is a surrogate for energy constrained target detection performance. We analyze the performance of the proposed multistage ARAP approach and show that it can reduce mean search time with respect to ARAP for equivalent energy constrained detection performance. To illustrate the potential gains of MARAP, we simulate a moving target indicator (MTI) radar system and show that M-ARAP achieves an estimation performance gain of 7 dB and a 85% reduction in scan time as compared to an exhaustive search. This comes within 1 dB of the previously introduced ARAP algorithm at a fraction of its required scan time.

Impact: This work provides a policy for mulit-scale search for sparse targets under total resource constraints, such as available scan time and/or computational resources. The proposed policy enhance previous work by considering the case where the number of measurements are constrained as well.

Journal: IEEE Transactions on Signal Processing, May 2011. DOI: 10.1109/TSP.2011.2112353. Full text available here.

Submitted by Gregory Newstadt, Graduate Student Research Assistant, Electrical Engineering and Computer Science, College of Engineering. newstage@umich.edu

Thursday, June 14, 2012

Distributed Supervisory Controller Design for Battery Swapping Modularity in Plug-In Hybrid Electric vehicles

Abstract: A distributed supervisory controller is proposed to achieve battery component swapping modularity (CSM) for a plug-in hybrid electric vehicle (PHEV). The CSM permits the designer to distribute a part of the supervisory controller to the battery module such that the PHEV can use a range of batteries while providing corresponding optimal fuel economy. A novel feedback-based controller for the charge sustaining mode is proposed to facilitate distributed controller design for battery CSM. The method based on sensitivity analysis of the control signals with respect to the battery hardware parameter is introduced to define the controller distribution architecture. The distributed controller gains are obtained by solving a bilevel optimization problem using the collaborative optimization and the augmented Lagrangian decomposition methods. The simulation results demonstrate that the proposed distributed controller can achieve battery CSM without compromising fuel economy compared to the centralized control case.

Impact: The battery module (including battery controller) can be swapped (e.g., upgraded or used to handle product variants) without having to redesign and retune the supervisory controller for the PHEV power management.

Journal: ASME Journal of Dynamic Systems, Measurement and Control, July 2012. DOI: 10.1115/1.4006214.

Submitted by A. Galip Ulsoy, the C D Mote, Jr Distinguished University Professor of Mechanical Engineering and Professor of Mechanical Engineering, College of Engineering. ulsoy@umich.edu

Automotive Control Systems

Summary: This engineering textbook is designed to introduce advanced control systems for vehicles, including advanced automotive concepts and the next generation of vehicles for ITS. For each automotive control problem considered, the authors emphasize the physics and underlying principles behind the control system concept and design. This is an exciting and rapidly developing field for which many articles and reports exist but no modern unifying text. An extensive list of references is provided at the end of each chapter for all the topics covered. It is currently the only textbook, including problems and examples, that that covers and integrates the topics of automotive powertrain control, vehicle control, and intelligent transportation systems. The emphasis is on fundamental concepts and methods for automotive control systems, rather than the rapidly changing specific technologies. Many of the text examples, as well as the end-of-chapter problems, require the use of MATLAB and/or SIMULINK.

Impact: It is currently the only textbook, including problems and examples, that that covers and integrates the topics of automotive powertrain control, vehicle control, and intelligent transportation systems.

Publisher: Cambridge University Press, June 2012. ISBN:9781107010116.

Submitted by A. Galip Ulsoy, the C D Mote, Jr Distinguished University Professor of Mechanical Engineering and Professor of Mechanical Engineering, College of Engineering. ulsoy@umich.edu