Visiting scientist research collaborations
Vibration reduction in a household appliance
Vibration control is important in many household products, automobiles, aircraft, and medical devices. Recently, active electric vibration control technology has been under development as a means to further improve vibration control.
In this research, a MIMO (Multiple Input Multiple Output) system model of a household appliance with active vibration control devices was identified by means of MATLAB/ident. The model was obtained from experimental data sets from a prototype household appliance. Controllers of the active electric vibration suppressors were designed by the Linear Quadratic Regulator (LQR) method. The controllers performed well in simulations. Research will continue in an effort to bring this to use in a real commercial product.
Manufacturing machines are getting faster to increase compatitivenes. Although faster speeds increase productivity, they also cause vibrations. Traditionally, a machine’s weight was increased to reduce vibration. However, because an increase of mass causes cost and time problems, vibrations are now being controlled by active and passive dampers. In this project, a tool was developed by which the value of the damper was optimally decided; the effects of the active controllers, optimal controller and H-inf controller were checked; and a simulation connecting ANSYS and MATLAB was developed which built a bridge between mechanical analyses and control simulators.
Control algorithms for autonomous rendezvous and docking with the SCAMP underwater vehicle
Autonomous rendezvous, docking, and proximity operations for spacecraft will require highly accurate, coupled position and orientation maneuvers. To provide the necessary performance, we investigated a family of passive, adaptive nonlinear control algorithms, and tested their performance in the Neutral Buoyancy Research Facility on the SCAMP underwater robot. We developed a detailed physical model for the dynamics of SCAMP, and used this to determine the necessary structure of the control algorithm. MATLAB simulation studies verified that the controller could achieve the desired accuracy on simulated maneuvers. We then ported the Matlab code to C and integrated it with the real-time flight control code on SCAMP. Underwater tests validated the performance of the new code in "real-world" conditions.
A novel parallel hybrid transmission
A parallel hybrid vehicle typically employs two or more power sources to drive the vehicle. For such a vehicle to function properly, a non-conventional transmission mechanism and a microprocessor-based controller are needed to manage the power flow among the various power sources. The transmission mechanism described in this paper can provide a parallel hybrid with thirteen clutching conditions that can be grouped into five major modes of operation, namely, electric motor mode, power mode, CVT/charging mode, engine mode, and regenerative braking mode. The kinematics, statics, and power flow of each mode of operation are analyzed. A numerical example is used to illustrate the principle of operation. Furthermore, a clutching sequence control logic is developed.
Microelectronics process technology
The ability to interconnect multiple chips at different elevations on a single substrate could significantly improve the performance of advanced small optical modules and reduce the package size of a MOSFET relay.
In their project, “Compact Packaging Using MEMS-Based 3-D Substrate Interconnects,” researchers developed compact 3-D silicon substrate interconnects. Gray-scale lithography technology was used to create smooth inclined surfaces between multiple vertical levels. Successful electrical interconnection was established via metal evaporation and contact lithography using spray coating.
Optimization-based available-to-promise with multi-stage resource availability
Customer service, rapid response to customer requirements, and flexibility to handle demand and supply uncertainties are becoming strategic differentiators in the marketplace. Organizations need sophisticated approaches to conduct order promising and fulfillment, especially in a high-mix, low-volume production environment.
The Available-to-Promise (ATP) function has migrated from a set of availability records in a Master Production Schedule (MPS) toward an advanced real-time decision support system that enhances decision responsiveness and quality in Assembly To Order (ATO) and Configuration To Order (CTO) environments. Advanced ATP models and systems directly link customer orders with available resources, including material and production capacity.
The research adapts previously published mixed-integer-programming (MIP) models to specific requirements posed by an electronic product supply chain within Toshiba Corporation. The model provides individual order delivery quantities and due dates, together with production schedules, for a batch of customer orders that arrive within a predefined interval. It considers multi-resource availability including manufacturing orders, production capability and production capacity. It also takes into account a variety of realistic order promising issues such as order splitting, model decomposition and resource expediting and de-expediting. The research was published in Springer’s Annals of Operations Research, Vol. 135, No. 1.
Hardware and algorithm development for control of bipedal locomotion
This research focused on the control of legged locomotion. The experimental setup was assembled with a commercially available humanoid robot platform connected to a Matlab/Simulink based controller. The main focus of the algorithm development was the procedure for designing virtual constraints for the legged robot with time-independent control that has a potential to improve the walking robustness against external disturbances. The virtual constraint design is based on Linear Inverted Pendulum mode that provides simple orbital energy calculation. The contents also include a discussion of the concept of energy based control and simulation based investigation of the effect of the feedback linearization.
Exhaust emission control
The researcher developed an engine idle speed and emission controller and compared it to existing controllers. Modeling was extensively used; the engine model included airflow dynamics, combustion, fuel injection and catalytic converter models. A linear model was developed by linearizing at nominal points. Measured and simulated data were compared to evaluate the accuracy of the model. The models were used to compare idle controllers, air/fuel ratio controllers, and emission controllers.
Optimization and simulation models for ground delay programs
This research determined the cost of three forms of flight delay uncertainty—flight cancellations, unexpectedly arriving flights and flights that deviate from their assigned arrival times—and the value of reducing them. It also looked at whether changes could be made to ground delay program (GDP) planning or execution to better mitigate demand uncertainties.
Optimization and simulation models were developed that generate effective planning strategies for a stochastic demand and deterministic capacity scenario. These models incorporate uncertainty in demand by associating probabilities with the stochastic demand elements during GDPs.
Such enhancements to ground delay management programs are a major component of Collaborative Decision Making (CDM), a potential new way of handling air travel that brings together airlines, government, private industry and academia and has the potential to reduce flight arrival delays in the United States. With CDM, airlines share their latest schedule information with each other and the Federal Aviation Administration (FAA) to improve air traffic management decision making.
Optical wireless tracking
A laser transmitter and receiver in an optical wireless communication link must face each other in a straight line to maintain the link. When either are misaligned, the link will terminate immediately. This has been one of the most significant problems of optical wireless systems.
To overcome it, an optical wireless tracking system that can accurately and rapidly align the transmitter and receiver is required. For real-time processing of positioning, the actuator requires motion at high speed.
The researcher designed a prototype optical wireless tracking system with two five-phase stepping motors controlled by a program on a personal computer. Actuators track the receiver’s position. The system is able to rotate 360 degrees horizontally and vertically in less than one second.
Data mining algorithms
Data mining algorithms are improving manufacturing process performance and efficiency. The joint project established a theoretical background and an improved data mining technique to analyze huge manufacturing databases. Researchers identified the need to identify defect causes, then improved manufacturing yield by reducing them. An algorithm was developed which was tested on a known data set and compared to existing techniques. The new algorithm consistently extracted the correct answer, executed more quickly, and was more robust than existing techniques.
Device Modeling and Characterization of 4H-SiC npn Power Bipolar Junction Transistor
Semiconductor power devices are key components to realize power converters with high power output and density. Silicon carbide (SiC) devices are expected to be the next generation of power converter devices. The wide bandgap and high thermal conductivity of SiC enables design of power devices that can operate at high power and high temperature. 4H-SiC bipolar junction transistors (BJTs) are one of the promising candidates for next-generation power devices. This research executed the modeling of the visiting scholar's company’s original BJT: the suppressed surface recombination structure (SSR)-BJT using a unique 4H-SiC device simulator, and then confirmed the static and the dynamic characteristics using this device modeling.
Collaborating with technical domain experts
Visiting scholars have worked with ISR and other University of Maryland faculty to learn specific technology topics that are being applied within their organizations. They have studied probability and statistics, reliability engineering and analysis, statistical process control (including change detection), and independent component analysis to support the extraction of hidden information in manufacturing data.
This work was achieved through attending university courses, working homework problems, meetings with faculty and students, extensive self-studying, and through case study analysis applications of these techniques.
Need more information?
For more information, contact Jeff Coriale, 301-405-6604.
Return to the ISR strategic partners main page.