DRONES Funded Research - 2016
" Application of amphibious drone technology to enable environmental monitoring and sampling to support disaster response and drinking water resource protection",
Dr. William Alexander, Dept. of Chemistry
Herein, we propose to investigate multiple aspects related to amphibious air-to-surface-water drones for water sampling applications. We will use our expertise in analytical spectroscopy to investigate the technical feasibility and drop-in capability of current sensor and analytical technologies for on-board drone implementation. The cost and benefit of developing and operating such drones will be examined in light of water testing in routine and emergency response situations, with aspects of rapid deployment in emergency situations to disparate regions leveraging FedEx's logistical infrastructure. Our involvement in the emerging field of disaster response science related to source water protection will inform these cost-benefit analyses. We will use our surface science expertise to examine and identify promising low-wetting surface materials to help overcome energy losses inherent to drone liftoff from surface waters due to cohesive forces between pontoon/landing gear and the liquid surface. Finally, we will design and build prototype analytical payload platforms for various aspects of water quality sampling. We believe this proposal has the potential to directly impact the regional community, utilities, and governmental agencies; yield "big data" streams from environmental water quality observation and analysis; and will substantially expand DRONES research portfolio into new territory.
DEEP LEARNING ENABLED NON-INVASIVE COGNITIVE INTERFACE: WHERE MACHINE MEETS THE MIND"
Dr. Mohammed Yeasin, Dept. of Electrical and Computer Engineering
We propose to develop a natural yet accurate cognitive state/events/activities enabled interface for machine (for example, drone and robot). The proposed approach will use off-the-shelf wireless EEG headset, which will record human brain activities. The cognitive interface module will turn those signals into actionable control/interface commands. We will use off-the-self (toy) robots and drone to demonstrate the superiority of the proposed approach. One of the challenges in modeling cognitive events and states from electroencephalogram (EEG) recording is finding "meaningful representations" that are invariant to inter- and intra-subject differences, as well as to inherent noise associated with such data. We plan to investigate a novel approach for learning such representations from multi-channel EEG time-series. To accomplish robust modeling of cognitive activities and mental states (for example, emotion, epistemic state of mind, cognitive states, mental load, evoked thoughts etc.) we have need to build a computational framework that will consists of (but are not limited to):
- i. Transform EEG activities (i.e., powers in alpha, beta, and theta band) into a sequence of multi-spectral images. This approach will preserve spatial information that is ignored in standard EEG analysis.
- ii. Learn efficient representations and hierarchy of abstraction in cognitive activities from the sequence of images. We propose to use a deep recurrent-convolutional network inspired by state-of-the-art video analysis technique to learn cognitive events/states. The proposed approach would be designed to preserve the spatial, spectral, and temporal structure and information. This is expected to help in finding features that are less sensitive to variations and distortions within each dimension. This is fundamentally different from classic feature engineering approach which need to be customized for every application.
- iii. Modeling of cognitive states in the context of controlling off-the-self drone and toy robot.
- iv. Develop protocol and building cognitive interface to control external devices.
- v. Establish bench mark to objectively evaluate the performance of the cognitive interface for machine.
The scope of this research is much broader and will require sustained external funding. The goal of this pilot study is to conduct research to generate preliminary results and build a collaborative team. This will help the PI to explore grants from the NSF and National Aeronautics and Space Administration (NASA).
"Enhancing Human Capabilities using Unmanned Systems and Drones"
Dr. John Hochstein, Mechanical Engineering
Mr. Robert Pap, Accurate Automation Company
This proposal will fund some new work based upon projects by Dr. Chuck Jorgensen with aircraft programs that Robert Pap did for NASA Hypersonic programs, which have evolved into our autonomous boats and aircraft programs. We will go back to the basic concepts of the NASA research and develop a couple of features needed in tomorrow's unmanned programs. We will first have Dr. Jorgensen come to visit and lecture at the University of Memphis on how important the man machine interface is. Dr. Jorgensen is retired in Sedona, Arizona. We will see how one of these Device and/or methods can be integrated into an unmanned system to enhance the human capability. We will select one of projects for follow on research and go after funding from NASA or iARPA. Each of these technologies to allow the unmanned system to operate in a more robust way. We will look at safe operation of automation using these concepts that effect i. operator safety (C.C. Jorgensen, and S. Johan, 2013 ), ii. operator workload (C.C. Jorgensen and S. Johan, , and V. Lu, 2013 ) , and iii. Machine Oversight (Development of New Heartbeat Biometrics Interface (V. Khizhnichenko and C. C.Jorgensen, 2013) in a real world use. We will be consider the implications on an automation system like the Accurate Automation Sentinel unmanned boat and potentially the AAC GLOV UAV.
"Considering the Potential Impact of Autonomous Vehicles on Transportation Planning and Equity in Memphis"
Dr. Charles Santo, Dept. of City and Regional Planning
I propose to contribute to the CRAVD initiative by conducting research on the potential impact of autonomous vehicles ("driverless cars") on transportation planning and on urban form in Memphis. This work will have a particular emphasis on community engagement, bringing University resources and partnerships to bear in positioning Memphis to be proactive about the inevitable impact of this technology, so that it might be used to solve current urban ills and not exacerbate them. The seed funding of the CRAVD initiative would support outcomes related to 1) community engagement and impact, 2) scholarly publication, and 3) the development of trans-disciplinary Partnerships Inside and Outside of the University.
1. Community Engagement of Problem Solving
Memphis struggles with crippling poverty and intraregional inequity – issues that are inextricably linked to the region's sprawling urban form and the related ineffective public transit system. This research would explore whether and how driverless cars could improve equity within our city by addressing a transportation gap. A lack of emphasis on urban planning over the past four decades has contributed to existing disparities in the region. This research would help ensure that driverless car technology does not continue this trend.
2. Contributions to the Literature
This research would also address an existing gap in the literature. While driverless car technology is gaining attention in mass media outlets like Popular Mechanics, NPR, and the New York Times, very little related research has appeared within the discipline of city planning – a field that holistically considers the relationship between the built environment, land use, transportation, economic development and quality of life.
3. Building Trans-disciplinary Partnerships Inside and Outside of the University
This research also provides an opportunity to create new, and build on existing, partnerships between experts in multiple disciplines. The new University of Memphis Design Collaborative is currently preparing for a Workforce to Work Transportation Summit and is working in partnership with the City's Transportation and Mobility Project Manager and the University's Intermodal Freight Transportation Institute. The research proposed here could augment this ongoing work. In addition, the research would present opportunities to involve University and local corporate experts in supply chain management to problems related to routing and the logistics of driverless cars. Memphis is home to the world's foremost experts in logistics and moving cargo; the right partnerships can bring that knowledge to bear on applying new technologies to solve the problem of moving people.
"Secure Information Sharing among Autonomous Vehicles"
Dr. Lan Wang, Dept. of Computer Science
Autonomous vehicles are the future of transportation and it has been the topic of much recent research. An array of detection technologies, including sonar devices, stereo camera, lasers, and radars, can be used to acquire real-time information surrounding an autonomous vehicle, but whether the driving will be uneventful depends on the accuracy of the information gathered by the sensors. One obvious problem with sensors is their inability to acquire long-range data due to limited range and ﬁeld of view. Moreover, in some situations sensors can have diﬃculty acquiring short-range data, e.g., if the sensors are obstructed, they cannot sense what is beyond the obstruction. However, this problem can be solved if the vehicles can pass traﬃc information to each other. The implication of this approach is limitless: a vehicle which is miles far from a place will be able to know about an accident occurred in that place from another vehicle and change route accordingly. Cellular links, short-range wireless links, and roadside units are commonly used to connect vehicles to the network infrastructure or with other vehicles for information sharing. However, in most cases the information sharing is limited to one-hop and used for preventing collision. Multi-hop information sharing will enable vehicles to know the distant road condition and plan ahead. Vehicles will acquire information about their surroundings using local sensors and will share the acquired information with other vehicles. The sensors, however, will generate a plethora of data, so it is important for the vehicles to retrieve only the data that will help their decision making. Furthermore, since vehicles can move very fast, data needs to be distributed under transient connectivity. Finally, the outcome of exploiting security vulnerabilities is very serious for an autonomous-vehicular network, as false information may cause accidents. To summarize, challenges related to autonomous vehicular networks are transient connectivity due to mobility, large amount of time sensitive data, and potentially false information. In the next section, we will discuss how data-centric networking can address these issues. We propose to develop a prototype data sharing system for autonomous driving that will run over Named Data Network (NDN) , a new data-centric network architecture. If the proposed project is successful, the methodology developed in this project will become the basis of our research proposals to federal agencies. This research area is highly interdisciplinary, with collaboration opportunities with researchers in Transportation, Mechanical Engineering, Electrical Engineering, as well as Business Information and Technology.
"Legal Aspects of Drones"
Dr. Larry Moore, School of Accountancy
The Federal Express drone project provides me with an opportunity to continue to research
in an area of developing aviation law. I was fortunate to begin research in international
air law during the transition from the original Warsaw Convention that governed all
aviation since 1926 to the new Montreal Convention which was drafted in 1996 and to
which most major airlines and industrial nations transitioned. Being on the cutting
edge lead to multiple A+ journal publications. A national academy award for best research
paper of the year and the Fogelman College award for best research paper of the year.
One of the publications was in McGill's journal which was listed as the number 2 law
school and journal out of over 2000 law schools in the U.S. and Canada. At one time
I was listed as the leading researcher in this area in American and as one of the
leaders world wide.
I see the developing drone law as being strategically placed for purposes of academic research as was international aviation law when I began publishing in that area 25 years ago and with the same opportunity to be a leader in this area and to have some practical influence in the direction that this law takes.
But as was my research in international aviation, I want to be able to provide concrete guidance and insight into this developing area of law, such that companies such as Federal Express can operated with added clarity as to  the boundaries of regulations,  potential areas of liability both governmental and private, and  areas open to further development and expansion.
I will conduct this research by a careful review of the current literature and take what is now known so that when I review the voluminous new regulations currently proposed, I will be able to
focus my research on areas that are of interest to be explored academically or could be problem areas to avoid by such users of drones such as Federal Express.
"Engineering Novel light weight Supercapacitor: Batteries for Ultra-Light-Vehicles"
Dr. Sanjay Mishra, Dept. of Physics and Material Science
The proposed project focuses on developing novel light supercapacitor batteries for futuristic energy applications related to ultra-light vehicles including drones. The main thrust of the research is to develop novel nanostructured based Supercapacitor batteries which includes (1) identification of potential materials, (2) development of novel nanostructures, and (3) understanding of long term repeatability and stability of supercapacitor materials. The energy dense supercapacitors will allow effective long distance mobility of drones at appreciably low cost. The long term implication is that proposed efforts will result in discovery and engineering of marketable, energy dense, small carbon foot print, cheaper, material for transportation and energy applications. The research efforts will be interwoven with outreach activities including regular seminars, presentations at conferences, and high school student participation in various aspects of the project via on-going summer programs on campus.
"Restoring Damaged Metallic Parts of Robots, Autonomous Vehicles, and Drones by Additive Manufacturing"
Dr. Ebrahim Asadi, Dept. of Mechanical Engineering
The objective of this proposal is to establish a multiphysics computational framework and tool to predict the residual stresses and temperature map in the damaged metallic key-parts of robots autonomous vehicles, and drones (RAVD) that are rapidly restored using Additive Manufacturing (AM) without disassembling the parts from the system. RAVD consists of several metallic key-parts, often times with a precise and complex geometry, which are essential for the function of the system. In the case that these key-parts are damaged (either due to accidents or fatigue), it is desired to restore them in a short period of time to bring the system back to the ordinary function and minimize the related financial lost. The typical restoring process of the damaged parts includes two timely steps: 1) disassembling of the damaged part and 2) sending the damaged part to the manufacturer for fixing or replacing because it is not feasible to keep many spare parts at the warehouse. The proposed research will accelerate restoring of the damaged metallic key-parts of RAVD by repairing the damaged parts without disassembling them. A three-dimensional (3-D) image of the damaged section will be generated using cross-sectional scanning (CT-Scan) and an AM technique (called laser engineered net-shaping, LENS) will be used to fix the damaged section; the proposed approach is schematically shown at Figure (1). Although the proposed research will focus on utilizing LENS for restoring the damaged metallic parts for RAVD, LENS can be also used for on-demand and customized manufacturing of the whole parts directly from the digital blue-prints; thus, the proposed research will enhance the utilization of LENS for this purpose too. Therefore, the outcome of this research will directly benefit the local industry using RAVD and corresponding scientific communities.
AM for metallic parts (laser sintering) is a manufacturing technique to fabricate functional parts by layer-by-layer scanning of fine metal powders and melting (or partial melting) them using a high energy laser at a prescribed scan pattern [1-3]. The capability of AM for rapid manufacturing of complex-shape and fully functional parts directly from metal powders, without using any intermediate binders nor rough post-processing, is such an exciting possibility that cannot be matched by any traditional manufacturing technologies. The sale of AM products are currently $3.7 billion worldwide and it is foreseen to reach $6.5 billion worldwide by 2019 . Most of the AM techniques are based on sintering of the designed pattern from a pre-deposited rectangular thin layer of powders in a confined chamber that makes them inappropriate for fixing of the damaged parts without disassembling them; e.g. selective laser melting (SLM), selective laser sintering (SLS), direct metal laser sintering (DMLS), direct metal deposition (DPD), electron beam melting (EBM), laser powder deposition, and selective laser cladding (SLC) . In the other hand, LENS can be used for fixing the damaged parts without disassembling them because LENS uses Nozzles to spray powders to the desired location where a focused laser beam creates a melt-pool to fuse the powders to the substrate and to each other. In addition, different materials can be deposited by using couple of nozzles each one depositing different powders.
As it is apparent from LENS process, a cyclic heating and cooling occurs around the melt-pool and as a result in the part and, also, the whole RAVD system. The cyclic thermal treatment may cause: 1) overheating of other parts of the RAVD that do not have sufficient thermal resistance and 2) residual thermal stresses that may cause the catastrophic failure of the system at the next working cycles. Therefore, it is extremely important to predict the temperature map and residual stresses during LENS process and after that. We propose to develop a multiphysics finite element method (FEA) that combines heat transfer, mechanical equilibrium (including solid-liquid transitions), and Maxwell's wave equation (for modeling laser) for this purpose. The proposed computational framework and tool will enable the prediction of the temperature map and residual stresses to explore the feasibility of using LENS process for rapid repairing of the damages in specific RAVD systems; thus developing proper standards for LENS use in RAVD systems.
"Investigating the impact of adopting/using drones on the bargaining power of farmers in a contract process."
Dr. Euntae "Ted" Lee, Dept. of Business Information Technology
Dr. Hyungchul "Kevin" Kim, School of Accountancy
In general, agriculture business is very risky and unpredictable compared to other types of business because of natural disasters and uncertain weather conditions (Ahsan 2011). This high inherent risk involved in agriculture business can often force many farmers to make a risk averse choice (Hardaker et al. 2004). Moreover, due to high risk/ opportunity cost of selling their crops to customers directly, many farmers tend to choose to sell their products to a middleman or big company. As such, middlemen and big companies usually have high bargaining power in a contract process because of the low concentration ratio between middleman (and big broker company) and farmer, high degree of dependency upon existing distribution channels and their big volume purchase. This problematic structure of the agriculture business can often result in low profitability to farmers. The problem of low profitability in farming can in turn lead to severe financial issues in running agriculture business, and thus many farmers are gravely concerned with this problem of low profitability. In the hope of alleviating this problem of the low profitability in farming, we will investigate the impact of adopting/using drones or UAVs (Unmanned Aerial Vehicle) on the bargaining power of farmers in a contract process.
"Integrated Platforms and Algorithms of Multisensory Data Capture and Decision Support for Autonomous Vehicles"
Dr. Robert Kozma, Dept. of Mathematical Sciences
The present proposal addresses several key components of the FIT CRAVD initiative
in the area of autonomous navigation and control of vehicles and drones
using integrated platforms and algorithms of multisensory data capture and decision
support in dynamically changing, complex environments. The proposal benefits
from Dr. Kozma's proven past achievements in the field of intelligent robotics and
data fusion, including extensive collaborations with NASA JPL Robotics, and with Air
Force Research Laboratory, Sensors Directorate.
This proposal supplements and extends ongoing externally funded research on strategy change in complex environments, supported by NSF Collaborative Research in Computational Neuroscience. Moreover, this project will provide seeding support for upcoming large-‐scale, multi-‐disciplinary proposals on design environments for novel autonomous and manned land, sea or air assets and distribution systems (e.g., for DARPA). Example of such complex system in civilian domain is a resilient urban infrastructure assuring public safety both in normal and emergency scenarios, with reliable logistics, transportation, communication, and power. The project provides the opportunity for the involvement of local and regional partners, including FedEx, MLGW, benefiting economic growth, security, and social development in Memphis, TN, and the Mid-‐South region.