ESE Senior Design, 2011-2012
Ken Laker, Peter Scott
ESE Home Page > ESE Undergraduate Labs > Senior Design > 2011-2012 Abstracts
Cash for Trash: Prototype Planning Tool
for MBT Type Waste Management Facility
Authors: Divyang Arora, Kerimhan Nahum, Tibet Ors
Advisor: Dr. Peter Scott, Mr. Mehmet Ali Nalcacioglu and Mr. Richard Reinhart
Abstract:
A Mechanical Biological Treatment (MBT) waste management facility combines the separation of inorganic waste, and the biological treatment of organic waste yielding methane and compost. Separated inorganic waste can be utilized for recycling; the compost for landscaping or fertilizer production; and the methane to generate heat and electricity.
Revenues and costs relevant to these end products show drastic variability depending on the type and amount of waste processed by the facility. This variability makes long-term capacity and financial planning very challenging and merits the need for a user friendly tool: a tool specifically designed for an MBT type facility, which can accurately estimate the revenues, operational costs and total investment cost as a function of the mass and types of wastes to be processed at a particular location.
This tool addresses this need by forecasting the waste amounts to be processed by the facility throughout its lifetime using a regression analysis of past waste data. It then uses the forecasted amount of different waste types to estimate revenues and operational costs and determines the required total capacity and investment costs. It also enables the data generated to dynamically respond to various user inputs and graphically represents the generated results for effective interpretation.
View: Poster / Movie
mhsSIM: Mental Health System
Simulator
Authors: Justin Warner, Linfu Zhang, Zuri D Michan
Advisor: Dr. Barry Silverman and
Prof. Nancy Hanrahan
Abstract:
The current system of treatment for seriously mentally-ill (SMI) patients in the U.S. is plagued by problems of inefficient resource allocation and ineffective treatment procedures. This has resulted in high costs and poor success rates in the treatment of low-income mental health patients. In the Philadelphia community alone, this problem affects over 500,000 people and has resulted in substantial annual expenditures exceeding $850 million, ultimately funded by taxpayers. While researchers have proposed various policy solutions, they are currently unable to validate or quantify the effectiveness of their ideas. Meanwhile, healthcare officials and politicians are reluctant to make meaningful changes due to the high level of risk and uncertainty.
In response to this problem, our team has built mhsSIM, a tool using Agent-Based Modeling to simulate the dynamic interaction of the various agents and institutions in the mental healthcare system. In conjunction with researchers from Penn’s School of Nursing, we have developed a prototypical model that utilizes well-researched behavioral frameworks, extensive patient databases, and award-winning proprietary software to determine the impact of public policy on patient behavior.
Our tool allows researchers to experiment with policy changes in a risk-free, virtual environment. Initial results show that policies designed to reduce substance abuse, increase access to out-patient psychiatric nursing, and allow critical patients to remain in the hospital longer can ultimately improve the cost and treatment effectiveness of SMI patient care. Over the next two years, federal and private grants will be used to expand the mhsSim team and fund software commercialization.
View: Poster / Movie
PVS: Pacemaker Verification System
Authors: Sriram Radhakrishnan, Varun Sampath, Shilpa Sarode
Advisor: Prof. Rahul Mangharam
Abstract:
There are currently about 3 million individuals in the world who depend on pacemakers, which are surgically implanted devices that maintain proper heart rhythm. Pacemakers, complex as they are, are programmed with tens of thousands of lines of code, and are therefore inevitably prone to bugs. It may not come as a surprise then that between 1990 and 2000, over 200,000 pacemakers were recalled due to software issues. The purpose of the Pacemaker Verification System, then, is to test pacemakers with the realism of clinical trials but without their risks, with the goal of finding software errors before patient implantation. Our system addresses this problem with a virtual heart that can be reconfigured to exhibit different arrhythmias. The virtual heart was built by using finite state machines to model the hearts signal propagation properties for these different arrhythmias. Using code generation tools, these models were then directly translated to an FPGA hardware implementation, ensuring that the behavior of the models was retained. This virtual heart is able to interface with pacemakers and react to them in a manner that mimics a real heart. It can both output heart beats to the pacemaker and react to paces from the pacemaker. A top-level user interface allows manufacturers, regulators, and cardiologists to view the results of this closed-loop, dynamic system. Such a system allows for interactions that no static heart model can provide, thus offering more robust testing methods and thereby saving lives.
View: Poster / Movie
High-speed, High-precision A2D Interface for Flexible, Active Electrode Arrays
Authors: Akhilesh A Nayak, Eric L Chen, Benjamin K Shyong
Advisor: Prof. Jan Van der Spiegel and Dr. Jonathan Viventi
Abstract:
Collecting EEG brain potential data is an essential activity in neuroscience research. Our advisor, Dr. Jonathan Viventi has developed a novel, flexible electrode array design with characteristics that are superior to conventional non-flexible designs. This flexible electrode array can be applied directly to the surface of the brain and conforms to the intracranial folds. The array can capture rich electrical data, including the micro-variations from epilepsy seizures, and capture the data on a macro scale, as the array can be extended and applied over a relatively large surface area of the brain. Existing products for collecting analog data synchronously from brain electrodes are severely limited in sampling frequency, sample resolution, and the number of channels available for simultaneous capture.
To address this problem, we have developed a proof-of-concept high-speed (1 MegaSample/second), high-precision (24 bit), analog to digital interface that can capture data from one channel of the electrode array. Our design includes a schematic and layout for a PCB that takes an electrode array channel as an input and outputs digital data to a connector for a digital data acquisition board. This design is extensible to multiple channels - multiple copies of the PCB can be tiled and stacked to allow for more than one channel. Our board is able to successfully capture and convert the analog electrode data to a digital representation for analysis.
View: Poster / Movie
Authors: Jorge L Gracia, Ramzi M Ghannam, Raymond M Harari
Advisor: Dr. Andrew Huemmler
Abstract:
The demand for diesel fuel has become a global problem that developed countries are seeking to mitigate any way they can. An increasingly common response is through investment in alternative, renewable energy such as solar power.
When fire trucks go out on call and park on site, the plethora of electrical equipment on-board requires the engine to stay on, or idle, for the duration of the call. This wastes fuel and deteriorates the environment. Cities are looking for ways to reduce their carbon footprint.
We have designed a system that assesses the physical, economic, and environmental feasibility of attaching solar panels on the roofs of fire trucks (fire engines) for the purpose of saving fuel and reducing carbon emissions. This system is scalable and transferable to any city. The system uses call times, fire truck specifications and geographic solar energy data to calculate how much fuel consumption and carbon emissions would be reduced by, and hence how much the city would save, if it invested in solar panels for the fire trucks. The system performs an economic analysis of costs and savings (of installing panels) to calculate payback period, and ultimately makes the go/no-go decision (on investing) for the city's fire department.
The goal of the project is to help the user decide on solar panel installation on fire trucks through a user-friendly interface.
View: Poster / Movie
Helicopter Aircraft Wielding Kinect
Authors: Kevin D Conley, Paul M Gurniak, Mattew T Hale, Theodore T Zhang
Advisor: Prof. Rahul Mangharam
Abstract:
For as long as there have been natural disasters, there have been coordinated search and rescue (SAR) efforts to locate survivors. These disasters often radically transform the environment in which they have occurred, such as an earthquake damaging or destroying a building. In such instances, SAR personnel do not have any a priori knowledge of the areas they will enter or the hazards they may face.
The present state of SAR requires that rescue personnel enter a disaster area and look for victims as they explore. Their attention is divided between the two-fold task of maintaining their own safety and trying to bring any victims that they find to safety. While technology and SAR strategies have evolved over time, the task of locating victims still requires that someone jeopardize his or her own safety and enter an unknown environment.
To help mitigate this risk, we present a quadrotor system capable of mapping an unknown environment for SAR personnel. The quadrotor system consists of an Intel Atom-based on-board computer equipped with a Microsoft Kinect. Standard color camera and infrared depth sensory data from the Microsoft Kinect are wirelessly transmitted by the Atom to a base station laptop, where a three-dimensional map of the quadrotor’s surroundings is constructed in real-time.
Our proof-of-concept prototype demonstrates the efficacy of using a teleoperated unmanned aerial vehicle (UAV) for SAR operations. It provides the user with real-time visuals from the quadrotor accompanied by a 3D map of the quadrotor’s environment. SAR personnel can use this map to plan rescue efforts while minimizing risks to their personal safety.
View: Poster / Movie
NeoNur BT: Wireless Data Transfer and Analysis for Neonatal Feeding Device
Authors: Aaron M Roth, Daniel A Fleischer, Mattew G Sternberg
Advisor: Dr. Jay Zemel, Dr. Barbara Medoff-Cooper
Abstract:
More than half a million premature births occur in the United States each year. Researchers at the Children's Hospital of Philadelphia (CHOP) and University of Pennsylvania School of Nursing are currently analyzing the feeding behaviors of premature babies and how they relate to management of care and disease. The NeoNur system has been developed to measure, process and store sucking pressures and breathing data. The current system requires connection to a desktop computer via a serial to USB cable for processing and display. Our redesigned NeoNur system incorporates a wireless serial module that enables data transfer to Bluetooth-enabled tablets computers. Additionally we developed a mobile application that allows the feeding data to be easily transferred from the NeoNur and displayed as a graph along with key data analysis metrics. This allows nurses to quickly and easily view and analyze the feeding data, and determine if disease symptoms are present. The NeoNur could be used in hospitals throughout the United States. This system also has future applicability in other first and third world countries. In using this device, nurses will have the information they need to help save many premature babies around the world.
View: Poster / Movie
Authors: Amrita Nag, Daniel G Sherman, Megan R Purzycki, Christine T Muller
Advisor: Dr. Peter Scott,
Mr. Donald Bravo
Abstract:
Penn’s MOD 7 plant supplies chilled water to the entire campus for its air-conditioning and other cooling needs. Currently at MOD 7, an operator selects the thermodynamic parameters of the supplied water based solely on qualitative experience. This results in volatility of the selected parameters and overconsumption of energy.
In order to realize the savings that result from systematic determination of these parameters, there must be an accurate forecasting algorithm that predicts the chilled water demand. This project has developed a model using screened historical data and regression analysis.
The model takes as inputs the expected dry-bulb temperature, relative humidity, and cloud cover over the next 24 hours. The necessary data is automatically extracted from the Internet. The outputs are water flow, temperature differential between supply and return, and energy consumption. MOD 7 operators can use these outputs as the inputs into their current control center.
This model can be used not only in determining the necessary plant outputs, but also as a tracking and monitoring tool. It forecasts energy consumption for the following day, allowing the plant operator to proactively plan for the costly effects associated with spikes in chilled water demand. The university can also use this system to understand expected performance when doing life-cycle cost analysis of the plant and making upgrades.
View: Poster / Movie
Authors: Sanchit Kumar, Shantenu Agarwal, Daniel Lipschutz, Nina Zhu
Advisor: Dr. Beth Winkelstein
Abstract:
The effectiveness of the medical response to natural disasters depends as much on the organizational capabilities of the emergency responders as it does on the speed of the response. However, with many first response teams still dependent on clipboard and paper systems to manage patient records, resources get allocated inefficiently, patients get lost or forgotten, distressed patients are not treated urgently, and doctors are forced to choose whom they treat based on factors as trivial as proximity, rather than need.
The V-Prime system is our solution for managing patient records and monitoring patient conditions. This system is built to operate over a 500 square meter area. Central to this system is a device given to critically injured patients that wirelessly transmits each patient’s heart rate and oxygen saturation over a 24 hour period. A database situated on a central base station is used for storing and managing information about patient histories, diagnoses, and treatments, all of which is organized using electronic patient records. Finally, an application accessible on any WiFi-enabled device gives doctors easy access to patient records saved on the database, so that they can be retrieved and updated anywhere on the field. The application’s functionality supports the device by helping doctors with identifying patients, scheduling operations, and following patient alerts. Overall, V-Prime gives a concrete, affordable, and flexible platform for assessing patient health and managing patient treatments in emergency response scenarios.
View: Poster / Movie
Authors: Maurico Aizenman, Brandon D Christie, Lindsay Vanderhoff
Advisor: Mr. Kunal Gupta
Abstract:
Often times when dining out, large parties have difficulties splitting and paying their bill. Some people tend to consume more than others and processing separate checks is time-consuming and inconvenient for the restaurant staff.
These issues stem from the rigidness of the paper bill checkout system used by restaurants today. PayFast solves these issues by replacing the paper check with a digitalized bill on a tablet device. The system allows diners to split and pay the bill in a variety of ways. Restaurants can process payments more easily and form a customer mailing list from the checkout process. While testing the system in a restaurant simulation, diners preferred using the PayFast system to the traditional paper bill and benefitted from quicker and easier checkouts. They were also able to split the bill fairly by paying for what each individual ordered.
After integrating PayFast, restaurants can increase sales from higher table-turnover and improved customer relationships through their mailing list. Diners gain flexibility and leave happier when given more options in paying for their bill. The simplicity and elegance of PayFast provides these benefits without detracting from the overall dining experience.
View: Poster / Movie
Authors: Laura C Durrance, Jefta K Jawi, Agatha L Oliveria
Advisor: Prof. Jorge Santiago
Abstract:
On a daily basis, energy is constantly being wasted in both large and small scales. In the U.S alone, 56% of energy produced is wasted, and 20% of that comes from households. In an average household, most, if not all, appliances within it waste energy. Individually, the contribution of each of these sources seems trivial, but when aggregated over many users and long periods of time, the unused energy is substantial. This is a problem that energy harvesting aims to address.
The purpose of this project is to combine the efforts of energy and heat scavenging systems in the context of a household. We will be focusing specifically on energy being wasted through the use of incandescent lamps, which are at most 10% efficient. The Energy Scavenging System consists of a solar and thermal module embedded in a lamp, as well as a battery charging circuit and USB charger.
The solar module consists of solar cells connected in parallel to form a solar panel arranged in a way to increase the output current. The thermal module consists of a thermopile circuit attached to a heat sink in order to increase the temperate gradient it is exposed to, therefore improving the current and voltage it generates. The harvested energy will be stored in a battery for subsequent use.
View: Poster / Movie
Authors: Edward B Chan, Sudeep K De, Bryan J Hsu, Joon T Kim
Advisor: Dr. Keith Weigelt
Abstract:
Microfinance institutions (MFI) currently serve the financial services needs of 100 million people who live developing countries. Many of these individuals live in communities with no access to traditional banking services such as loans and savings accounts and must rely on MFIs for these services. With an estimated 500 million people who can utilize MFIs there is an estimated $200 - $250 billion funding gap. The chief obstacle to bridging the funding gap is the low profitability of MFIs.
“Microfinance Support System” proposes the ground work for a software package that will aid MFIs around the world in making better decisions with regard to their loan portfolios to increase profitability. The system provides a unique service not found in the current offerings of commercial business intelligence tools and seeks to provide actionable decision analysis and metrics to large multinational MFI. The objective it to provide better information synthesis to these MFI’s to increase profitability and consequently attract additional investments.
The system consists of three components. First is a commodity pricing model. To date this model has been able to predict prices six months forward with a range of 80% - 89% accuracy. Second is a cash flow prediction model. Using the predicted commodity price, the cash flow model outputs a suggested loan payback schedule for loan candidates based on their business characteristics. Third is an ideal-basket model. By applying modern portfolio theory, the model outputs investment weights for loan portfolios to minimize risk while maximizing returns. Together these three models form a powerful toolbox which MFIs can rely upon to make better informed decisions.
View: Poster / Movie
Nigeria-Sustainable Population Growth Management
Authors: Jayme Chen, Seung-Jae Lee, Justin Y Yeh
Advisor: Dr. Peter Scott and Ms. Katherine Williams
Abstract:
On October 31st, 2011, the United Nations reported that the world population had reached 7 billion and was projected to hit 9 billion by 2050, and 10 billion by 2100. Crossing this benchmark has re-instigated awareness of the dangers of overpopulation. Scientists Donella and Dennis Meadows have updated their previous study in Limits to Growth to this century and have shown how the current world population growth rate is unsustainable. Nigeria is currently the fastest growing African country in population, and is predicted to be the world’s third most populous country by 2050.
There are many interdependent relationships between population growth and various aspects of the country. The purpose of this project is to develop a forecasting and decision-making tool that allows stakeholders to explore different prospects of sustainable development based on population changes. The model provides information on what directions the country is heading towards and how to modify that direction by influencing population change.
We have narrowed our model to focus on the central issues pertinent for Nigeria, including education and HIV. We used the system dynamics software STELLA to build a model that takes into account the various key factors affecting population growth and their impact on social, economic, and environmental aspects of Nigeria. We were able to compare the results of our model with existing historical data for validation.
Users such as policy makers, experts in population studies, and non-governmental organizations can see the potential impacts of population on social, economic, and environmental aspects of Nigeria.View: Poster / Movie
KENERGY: Modeling Location-Specific Renewable Energy in Kenya
Authors: Anne C Bancroft, Renee M Byers, Emily H Shaeffer
Advisor: Dr. Peter Scott and Mr. Peter Eschenbrenner
Abstract:
Much of rural Kenya is without adequate access to electricity. Rural Kenyans lack access to power grids, rely on kerosene lamps to study at night, and must make long trips to charging stations for their cell phones. Kenya has vast amounts of untapped renewable energy resources (hydropower, solar, wind, biogas). Presently numerous NGOs are interested in solving Kenya’s rural energy problem with renewable energy.
Kenergy is a tool designed to aid these NGOs in deciding which combination of renewable energy sources to pursue in a given location. Our linear model takes inputs of latitude, longitude, and either desired cost of initial investment or amount of energy provided. A database with the availability of each renewable energy source by location provides specific data. This data and the user inputs serve as constraints for linear optimization. The model is optimized to either provide the greatest amount of energy, when initial investment is input, or the lowest cost, when desired amount of energy is input.
Designing linear models for Kenya is challenging because there are many intangibles. With guidance from our stakeholders, we designed additional inputs for situational information to customize the results. These include externalities like environmental impact and lifestyle changes.
The final result is the specifications for a project in the location they designated. The user is given the final total cost and energy output of the project, and the breakdown of the four energy sources. The number, cost and type of each unit for each energy source is listed.
View: Poster /
Movie
Authors: Ruxin He, Yuanjiao Shen, Xiaoting Zheng, Ting T Zhou
Advisor: Dr. Andrew Huemmler
Abstract:
Penn Transit Services, a division of University of Pennsylvania, offers Penn Shuttle service that takes students and faculty from on-campus stops to their off-campus destinations within covered areas. With roughly 16 shuttles on the road at any given time (during service hours), Penn Shuttle delivers thousands of passengers every day.
Currently, as passengers get on the shuttle, the driver writes down the destinations on the log sheet. At the end of pick-up, the driver has to figure out the best route going through as many as 15 addresses. However, due to the high turnover rate of shuttle drivers, they typically lack the knowledge and experience required to ensure efficiency of the delivery process. Extensive efforts are spent in training the drivers, and the quality remains inconsistent. In addition, the log sheet information has to be manually typed in to the database at the end of every day, which is also a costly process.
PennRoute improves the current situation and overcomes the challenges by developing an integrated system implemented on Android tablets installed on the shuttles next to the drivers. The system generates the optimal route for the driver real time in the delivery process, records the log information automatically, and provides easy-to-navigate user interface.
PennRoute brings significant benefits to Penn Transit by reducing fuel, maintenance, training, and labor costs, and to the entire Penn community by improving the shuttle services and creating environmental benefits. The project is awarded Penn Green Fund Grant, and it is successfully deployed on four shuttles.
View: Poster / Movie
Authors: Valerie G Cerasuolo, Kara R Hollis, Michael R Simon, Harold J Stuckey
Advisor: Dr. Barry Silverman
Abstract:
MyDiaText is a text message reminder service designed to improve the effectiveness of the current diabetes education program at the Children’s Hospital of Philadelphia’s Diabetes Center for Children. Type 1Diabetes is the second most prevalent chronic disease of young adolescents. The lifestyle and behavior changes necessary for proper diabetes management are a significant challenge for this population. Working with the Diabetes Center for Children, MyDiaText uses popular texting technology to assist newly diagnosed diabetes patients in adapting to this lifelong disease.
With the growing population of adolescents with diabetes, this technology-based intervention may prove more effective than traditional classroom diabetes education alone. Using the online portal, www.MyDiaText.com, newly diagnosed patients choose a behavioral goal they wish to modify. These goals range from counting carbohydrates to establishing healthy exercise habits. Patients receive text messages throughout the program as a reminder to follow their chosen goal and are also prompted to rate their progress. As patients respond and receive reminder text messages, all activity is logged to their MyDiaText profile. Patients and their families use the online portal to track progress throughout the one-month program. The website also hosts educational quizzes and trivia to provide further diabetes education.
The Diabetes Center for Children hopes to pilot a one-month trial of MyDiaText to their patients in May 2012. Once the pilot is completed, participants will be asked to complete a patient satisfaction survey. Responses to the survey will help assess if MyDiaText was helpful to patients and the feedback will help improve MyDiaText for future use.
View: Poster / Movie
Authors: Aditya V Jayanthi, Madhav M Nandipati, Khaled Saad, Arshan B Vakil
Advisor: Prof. Jan Van der Spiegel
Abstract:
Course scheduling is an integral part of the college experience. At the University of Pennsylvania, students have the opportunity to choose among 600 courses across 120 departments. Although the breadth and depth of course offerings is a true benefit of the university, students face difficulty in scheduling their courses. The existing tools are not user-friendly, and information must be manually gathered from various online applications in order for students to make decisions about which courses they want to take.
The objective of this project is to integrate the various course scheduling tools at the University of Pennsylvania with an interactive and easy-to-use interface. Time efficiency is also an important goal, and with this single platform, course scheduling is a faster and more convenient process.
The approach of this project is to develop a web-based application. This system allows students to input their preferences by responding to successive queries. With these constraints, the scheduling program draws information from the Course Registrar and Penn Course Review databases to generate a ranked set of schedules.
View: Poster/Movie
Authors: Karan M Dhruve, Aditya Khosla
Advisor: Dr. John Pourdehnad
Abstract:
Food Trucks have been a landmark at the University of Pennsylvania for many years. There are 59 food trucks on Penn’s Campus today, all of which are well known for breakfast or lunch. However, almost anyone who has ever eaten at a food truck has experienced a substantial wait time. Since most students do not have much time for lunch in between classes, they usually cannot eat at food carts with long lines. Although some truck vendors have allowed for different ways to order, such as phoning in or texting an order, this only applies to a select group of food trucks.
PennTrucks.com is taking an approach to reduce this disconnect between customers and vendors It is an online application that facilitates ordering from food trucks on Penn’s campus. It is made available on smartphones and computers. The overall purpose of this project is to develop an application that streamlines food ordering from food trucks on Penn’s campus, minimizes the waiting time for customers and provides added benefits to food trucks.
After identifying the needs of the stakeholders, a website was created that allows customers to place orders and vendors to receive them. A market test of selected customers revealed that the average waiting time decreased from 8.6 minutes to 4.8 minutes and 90% of the respondents said they would definitely use PennTrucks.com
View: Poster / Movie
Authors: Naveen Kasthuri, Evan Hodges-Le Claire, Andrew T Piskai
Advisors: Dr. Andrew Huemmler
Abstract:
The Arbitrage Garage system simultaneously addresses two problems: high institutional electricity costs, and barriers preventing widespread electric vehicle adoption. With vehicle-to-grid (V2G) technology, electric car batteries can store cheaply purchased power to be discharged at times when the power would be more expensive to obtain. By utilizing excess electric car battery capacity, electricity cost savings are realized through reducing peak-load contribution (PLC) costs, and by performing arbitrage in hourly electricity spot markets. PLC costs account for one fifth of the University of Pennsylvania’s (Penn) annual electricity charges, yet are assessed during only five hours over the entire year. The Arbitrage Garage software controls electric car charging and discharging to maximize PLC and arbitrage savings, while always providing enough charge at all times for drivers to make their journeys home. V2G applications with private fleets of cars are well established because the times when the cars are parked are highly controllable; however, electric cars owned by individuals had previously been largely unexplored. Although the Arbitrage Garage software is the centerpiece of the system, this project explores everything needed to implement the software. This includes an economic analysis of a phased introduction of transmission line upgrades and individual DC fast charger units for each space in “Garage 37” located at the intersection of 34th and Chestnut streets. Each parked car saves Penn $1,600 annually, with each driver receiving about $500 just to park their car. The internal rate of return of the system’s 30-year study period is impressively well over 20%.
View: Poster / Movie
Authors: Timothy A Hennelly, Sarah Clark (MEAM), Sydney Jackopin (MEAM), Jacob Orloff (MEAM), Michael Posner (MEAM), Michael Siegel (MEAM)
Advisor: Dr. Bruce Kothmann
Abstract:
In the event of a natural disaster, communication and transportation networks are often disabled, rendering victims stranded and in need of supplies. Similarly, remote military outposts are hard to reach but frequently find themselves in need of resupply.
The current solution to both of these problems is to deliver supplies via helicopter. These resupply missions are dangerous and expose air crews to the potential for injury and death. Unmanned aerial vehicles (UAVs) provide one option for taking humans out of harm’s way, but are currently only used for surveillance and attack missions.
The system presented uses a UAV to deliver supplies to a target location via a parachute. The UAV’s guidance system uses GPS technology to autonomously navigate to a target location. Once on target, the system combines data from its sensor package with a drop trajectory model to calculate a release point for its payload. After releasing its payload, the UAV then autonomously flies back to its home base for human-controlled landing.
Dynamic models of the UAV’s physical specifications and the autonomous control scheme were created and used to develop initial parameters for the control code. By performing test flights and logging flight data, these models were validated and were used to build more complex autonomous behaviors. Stabilized flight, autonomous drop trajectory calculations with payload release, and autonomous navigation methods were all implemented and successfully tested.
.
View: Poster / Movie
Authors: Mattew R Kaye, Nithin A Puri
Advisor: Dr. Alejandro Ribeiro
Abstract:
The University of Pennsylvania supplies various types of physical resources to improve students’ learning environment, but unfortunately, these resources are not always being used efficiently. Some studying resources such as computer labs, study areas with desks, and private study rooms are under-utilized, while others suffer from overcrowding. Students do not have the knowledge or time to research these locations and as a result lose precious time searching for open studying space.
ResourceFull proposes a solution. ResourceFull monitors several study spaces, distributes information on availability to users through a web application, and enhances the overall studying experience for members of the Penn community.
View: Poster/Movie
ESE Home Page > ESE Undergraduate Labs > Senior Design Home Page >
2011-2012 Abstracts
ŠUniversity of Pennsylvania, Department of
Electrical and Systems Engineering.
This site maintained and administered by
Siddharth Deliwala, (deliwala_at_ee.upenn.edu ),
Last updated Apr 14, 2012