• 8 vertical mills with Proto-Trak 2 Axis CNC control
• 7 toolroom lathes with DRO
• 2 Okuma Genos L250 2 axis CNC lathe 
• 1 Okuma Spaceturn 3 axis CNC lathe
• 1 Romi 2 axis CNC lathe
• 1 Proto-Trak 2 axis CNC lathe
• 1 Okuma Millac 3 axis CNC mill
• 1 EMCO E350 3 axis CNC mill
• 1 Hardinge V480 3 axis CNC mill
• 1 Proto-Trak 3 axis CNC mill
• 1 Starrett HB 400 optical comparator
• 1 Zeiss CMM
• 1 Clasuing surface grinder
• 1 Marvel vertical band saw
• 1 Do-All vertical band saw
• 1 Willis radial arm drill press
• 1 Accutex  ZNC sinker EDM
• 1 Accutex 5 axis wire EDM
• 1 Rockwell Drill Press

Machine learning algorithms can sometimes do a better job with a little help from human expertise, at least in the field of materials science.

In many specialized areas of science, engineering and medicine, researchers are turning to machine learning algorithms to analyze data sets that have grown much too large for humans to understand. In materials science, success with this effort could accelerate the design of next-generation advanced functional materials, where development now depends on old-fashioned trial-and-error.

By themselves, however, data analytics techniques borrowed from other research areas often fail to provide the insights needed to help materials scientists and engineers choose which of many variables to adjust — and can’t account for dramatic changes such as the introduction of a new chemical compound into the process. In some complex materials such as ferroelectrics, as many as 10 different factors can affect the properties of the resulting product.

In a paper published this week in the journal NPJ Computational Materials, researchers explain how to give the machines an edge at solving the challenge by intelligently organizing the data to be analyzed based on human knowledge of what factors are likely to be important and related. Known as dimensional stacking, the technique shows that human experience still has a role to play in the age of machine intelligence.

The research was sponsored by the National Science Foundation and the Defense Threat Reduction Agency, as well as the Swiss National Science Foundation. Measurements were performed, in part, at the Oak Ridge National Laboratory in Oak Ridge, Tennessee.

“When your machine accepts strings of data, it really does matter how you are putting those strings together,” said Nazanin Bassiri-Gharb, the paper’s corresponding author and a professor in the George W. Woodruff School of Mechanical Engineering at the Georgia Institute of Technology. “We must be mindful that the organization of data before it goes to the algorithm makes a difference. If you don’t plug the information in correctly, you will get a result that isn’t necessarily correlated with the reality of the physics and chemistry that govern the materials.”

Bassiri-Gharb works on ferroelectrics, crystalline materials that exhibit spontaneous electrical polarizations switchable by an external electric field. Widely used for their piezoelectric properties — which allow electrical inputs to generate mechanical outputs, and mechanical motion to generate electrical voltages — their chemical formulas are usually complicated, including lead, manganese, niobium, oxygen, titanium, indium, bismuth and other elements.

Researchers, who have been working for decades to improve the materials, would like to develop advanced ferroelectrics that don’t include lead. But trial-and-error design techniques haven’t led to major breakthroughs, and she is not alone in wanting a more direct approach — one that could also more rapidly lead to improvements in other functional materials used in microelectronics, batteries, optoelectronic systems and other critical research fields.

“For materials science, things get really complicated, especially with the functional materials,” said Bassiri-Gharb. “As materials scientists, it’s very difficult to design the materials if we don’t understand why a response is increased. We have learned that the functionalities are not compartmentalized. They are interrelated among many properties of the material.”

The technique described in the paper involves a preprocessing step in which the large data sets are organized according to physical or chemical properties that make sense to material scientists.

“As a scientist or engineer, you have an idea whether or not there are physical or chemical correlations,” she explained. “You have to be cognizant of what kind of correlations could exist. The way you stack your data to be analyzed would have implications with respect to the physical or chemical correlations. If you do this correctly, you can get more information from any data analytics approach you might be using.”

To test the techniques, Bassiri-Gharb and collaborators Lee Griffin, Iaroslav Gaponenko, and Shujun Zhang tested samples of relaxor-ferroelectric materials used in advanced ultrasonic imaging equipment. Griffin, a Georgia Tech graduate research assistant and the paper’s co-first author, did the experimental measurements. Zhang, a researcher at the University of Wollongong in Australia, provided samples for the study. Bassiri-Gharb and Gaponenko, a research affiliate in her group, developed the approach.

A single crystal sample is placed onto the measurement stage of the modified atomic force microscope (i.e. piezoresponse force microscope). (Photo: Rob Felt, Georgia Tech)

Using a conductive tip on an atomic force microscope, they examined the electromechanical response from a series of chemically related samples, generating as many as 2,500 time- and voltage-dependent measurements on a grid of points established on each sample. The process generated hundreds of thousands of data points and provided a good test for the stacking approach, known technically as concatenation.

“Instead of just looking at the chemical composition that provides the highest response, we looked at a range of compositions and tried to figure out the commonality,” she said. “We figured out that if we applied this data stacking with some thought process behind it, we could learn more about these interesting materials.”

Among their findings: Though the material is a single crystal, the functional response showed highly disordered behavior, reminiscent of a fully disordered material like glass. “This glassy behavior really is unexpectedly persisting beyond a small percentage of the material compositions,” said Bassiri-Gharb. “It is persisting across all of the compositions that we have looked at.”

She hopes the technique will ultimately lead to information that will improve many materials and their functionalities. Knowing which chemicals need to be included could allow the materials scientists to move to the next phase — working with chemists to put the right atoms in the right places.

“The big goal for any materials’ functionality is to find the guidelines that will provide the properties we want,” she said. “We want to find the straight path to the best compositions for the next generation of these materials.”

This research was supported by the National Science Foundation (NSF) through award DMR-1255379, by the Defense Threat Reduction Agency (DTRA) though grant HDTRA1-15-0035, by the Center for the Science and Technology of Advanced Materials and Interfaces (STAMI) at Georgia Tech, and Division II of the Swiss National Science Foundation under project 200021_178782. The piezo-response measurements were in part performed at the Center for Nanophase Materials Sciences at Oak Ridge National Laboratory, which is a U.S. Department of Energy Office of Science User Facility. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the sponsoring organizations.

CITATION: Lee A. Griffin, et al., “Smart machine learning or discovering meaningful physical and chemical contributions through dimensional stacking” (NPJ Computational Materials, 2019, https://rdcu.be/bOycU).

Ph. D. graduate David Montes De Oca Zapiain didn’t plan to be an entrepreneur, but a great experience in Georgia Tech’s TI:GER program encouraged him to turn his research into a business plan that he is pursuing while working as a postdoc at Sandia National Laboratory. In this Q&A he talks about how he ended up staying at Tech for three degrees, and how he found a surprising level of support for taking his idea to market.

Where are you from and how did you end up at Georgia Tech?

I was born and raised in Mexico City, Mexico. I went to school there all the way through high school and I first came to Georgia Tech to pursue an undergrad in Mechanical Engineering.

The main reason I decided to come to Georgia Tech, aside from it being a great school, was thanks to my granddad (the father of my mother) since he studied at Georgia Tech back in 1957-58. Therefore, I've always had a strong connection to Georgia Tech.

 In 2011 I applied and got accepted, three years later I graduated with my bachelor's degree in mechanical engineering. I then decided to pursue my master’s in mechanical engineering at Georgia Tech-Lorraine ni France. While I was over there, I was fortunate enough to meet Dr. Kalidindi since he was co-teaching a class called Micromechanics of Materials. At the end of the semester, he offered me a Ph.D. position in Atlanta and that is how I ended up back in Atlanta again.

What was the focus of your research as a Ph.D. student?

The main focus of my research is how to deploy new materials into the market faster. I address this challenge by optimizing the screening process of new materials.

My main research project consists of developing a computational tool that can enable engineers and material developers to optimize the rate at which they can perform crystal plasticity finite element  simulations. This type of simulation enables material developers to predict accurately how a metal will deform during a forming process without the need to perform capital intensive and time consuming experiments.

As a matter of fact, this research project is the basis for the company I launched through Georgia Tech’s TI:GER program. The vision we have for the company is to  be a robust standalone software simulation suite that enables material developers and material scientists to obtain accurate numerical simulations at the fraction of the computational cost, and 1,000 times faster.

What need is your business addressing?

One of the problems we're addressing is that it takes a lot of time and money to make a new material. As an example, it took Boeing more than 20 years and over $32 billion in funding to develop the alloy that they use in the 777. This long lead time is needed since the materials designers and engineers need to be very very sure of how the material will behave and perform under the operating conditions. Especially, if this material is going to be used in an airplane.

The most accurate  way to do that is to do physical experiments. Having said that, you can't build a new airplane every time you want to test something. For this reason, they leverage computational analysis which gives them insight into how the material will perform without the need to perform expensive physical experiments. However, at this point they face a typical engineering dilemma- do I want to wait for more accuracy, or do I want to obtain results quickly?

Because if you want to get a really accurate prediction of how the material will behave you have to incorporate more physics into your model and you're almost getting back to the same amount of time and cost as if you were doing a physical experiment.

So that actually leaves a really well defined market that has been under-served by the incumbents. The market wants results that are fast and also accurate. Our company will come in and sell results that are fast and accurate, and we will do it using our proprietary machine learning and data-driven solution.

How has the process played out for you to turn your research into a business plan?

I started working to commercialize my research through the Georgia Tech TI:GER program. They paired me with three MBA students from the Scheller College of Business. They were Matthew Davis, Raymond Chu, and Andrew Bunch. They also paired us with two JD students from the Emory Law School- Landon Smith and Tylor Espy. We worked on developing a business plan for three semesters.

During this process we identified the market and quantified how much money exists to be made. Then, we identified potential customers, through a comprehensive process of customer discovery, validation of hypotheses, and finally integrating everything into a concise business plan with a robust financial model that would enable us to get attention.

The lawyers from the group were able to secure jobs in law firms, and the MBA students have received other jobs. I'm currently working with Matthew Davis. He graduated from his MBA program in December and is working at GTRI. He and I are working together to get this running. The company is called SwiftMat LLC.

The TI:GER program was very helpful in helping us navigate all of that process. We decided to do an LLC because it offers you the most freedom and is the easiest to incorporate.

Who is your target market for this service?

We have talked to many people in the industry and we have some potential customers in mind thanks to the enormous support  of the Georgia Tech VentureLab since they funded us to attend the Materials Science and Technology conference in October of 2018. This conference hosts over 2000 material developers and more than 100 companies.

During this conference, we validated many of our business hypothesis and zeroed in on the business need we are addressing. The feedback obtained through this conference coupled with the market research we previously performed allowed us to identify three major markets, or customer segments that we are going to address.

One would be private industry, such as aerospace companies or automotive companies, as well as oil and gas and other energy companies.

The second customer segment we identified was higher education and research institutes.

Finally, another segment that was very interested in our approach was national labs, specifically the Department of Energy Labs like Sandia.

What is the next step for getting SwiftMat up and running?

The next step is to seek out grant and seed funding. We've been actively applying through the Georgia Research Alliance (GRA). It's a good amount of money and best of all it is non-equity since it is a grant, which is exactly the kind of funding we we're looking for at this stage. We're also applying to a lot of entrepreneurship and business case competitions.

One of the main things we're looking for right now is an injection of cash to fully commercialize our software since at this point it’s still a research code with a high degree of complexity and not as user friendly as it will need to be. Someone would need me to run the software.

With an injection of cash, we could turn it into a product. We envision sending it out as through a software as a service (SaaS) model in the cloud, or locally. That's what we're looking for and why we need funding.

What business competitions did you participate in this year?

We participated in the TYE Atlanta Young Entrepreneurs business case competition in which 50 university startups competed. The preliminary round was in February and we advanced to the finals with the top eight teams. In the finals we presented a 10-minute pitch to an industry panel of four different judges. They were from industry, venture capital, and local startups in the Atlanta area. They wanted to see a demonstrated need for our business, the viability of our business, and our plan to monetize it. We were fortunate enough to land in second place, and that secured us $3000 in funding for the company. It also exposed us to the environment of entrepreneurship in Atlanta. We were given a scholarship to join their organization and be able to leverage their services for a year. They also have an investment side that we are exploring applying to.

Was that your first competition?

It was the first one where we had to deliver a pitch. We applied to three- Baylor's business competition where we made it to the top 20 or 30 teams, and the Rice University competition where we also finished in the top 20 or 30.
 

Did you envision launching a company when you came to Tech?

Not really. Not at all actually. For me it has been a huge opportunity. The opportunities that Georgia Tech has given me have been incredible. I'm so grateful. I came here as an undergrad and took advantage of everything that I could. I did the co-op program, internships, study abroad. I did a co-op with Toyota of North America my junior year. I interned with Nissan North America my senior summer. I studied abroad at Georgia Tech-Lorraine for a summer and I did the China summer program. Then I did my master's at Georgia Tech Lorraine. When I came back and started my Ph.D I had two more internship experiences- one with Sandia Labs and another one with Schlumberger in Texas. Georgia Tech has put me on many different paths and opened so many doors for me. The fact that I have had such a great school behind me has helped me a lot.
 

What were your priorities when you were looking for jobs after graduation?

One of the main things I looked for when I interviewed is how open the organizations are to entrepreneurship. When I interviewed with Sandia for a postdoctoral opportunity they were completely open. They offer a leave from work to pursue a startup, and if it works out it works out, and if it doesn’t, they are able to rehire you. So that's a great opportunity.

I want to push this forward. We're trying to get grants and seed funding. I was pleasantly surprised how many resources Georgia Tech has for new ventures. Specifically, the VentureLab and ATDC, and the Georgia Research Alliance.
 

If a current or future Ph.D. student was interested in commercializing their research what advice would you give them?

One of the main resources that every Ph.D. student should take a look at is the TI:GER program. The TI:GER program is one of the best initiatives Georgia Tech has because it allows you to get exposure to the business side. When you're getting your Ph.D. you're normally thinking about science, and you're not thinking about translating it into business. In my class there were seven different projects all the way from cancer cures to novel drug deployments to machine learning. They prepare you very well. Georgia Tech consistently places teams in first and second place in these big business competitions, and they all come from the TI:GER program. If any Ph.D. student thinks their research could have commercial impact they should look into it.

Editor’s note:Since this conversation, a provisional patent has been filed through the Georgia Tech Research corporation based on David’s research. 

Assistant professor Sourabh Saha joined the Woodruff School this summer and is still getting settled in as he adjusts to life on campus in the heart of Atlanta. In this Q&A he talks about his path to Atlanta, the motivations behind his research, and what drew him to Georgia Tech. 

Where are you from and how did you end up at Georgia Tech?

I'm originally from India and I did my undergrad at IIT Kanpur, then I moved to MIT for my PhD. All of my degrees are in mechanical engineering. For the last four years, I had been working at Lawrence Livermore National Lab, first as a postdoc and then as a member of the technical staff. I started here at Georgia Tech this summer. Being in academia was always my plan and I was searching for a place that values hands on engineering, fundamental science, and at the same time, is focused on making an impact on the world. I knew little about Georgia Tech when I interviewed here. And when I interviewed, it felt like the right place for me, so that's why I'm here.

Why did you choose a career in academia?

I chose a career in academia for two specific reasons. The first one is the type of research that I can do. The type of research that I can do here is focused on long-term impact and not just on short-term profit-making goals. Then there is the broad impact aspect, which is where the work is not just for a very specific mission, but is broadly applicable. And another important aspect is the intellectual freedom, which is being able to choose the problems that I want to work on. And that's not quite possible in industry or national labs. And the second part really is the training opportunities- being able to train the next generation of the workforce. Students here at Georgia Tech are very excited to learn new things. You wouldn't get that excitement in the industry workforce, or even with partners at national labs. It's exciting to work with that kind of people on a day in day out basis.

Who has influenced you in your career?

My research mentors, starting from my internship during undergrad, my Master's research advisor at IIT Kanpur and then my PhD advisor at MIT. All three of my research advisors had an influence on my career. In addition to that my parents had a strong influence on my career trajectory. My mom is a retired school teacher and my dad is an electrical engineer and I have now merged teaching and engineering in my career.

What are you teaching this semester?

 This semester I'm teaching ME 3210. It's a required undergraduate class in mechanical engineering and it's on manufacturing. The way I'm teaching it is for students to role play as real world engineers and to learn how to make good engineering decisions within the context of manufacturing.

What is your research area? My research is in the area of manufacturing, and specifically focusing on scaling of advanced manufacturing processes. It involves building tools and generating process knowledge to scale up the fabrication of micro and nano scale structures for use in real world applications.

Why did you get interested in this particular field?

So, I was always interested in making things and that got me into manufacturing for my graduate education at IIT Kanpur back in India, and even in my PhD program at MIT. But the focus of my current research, which is on scalability and making sure we have the manufacturing technology to bring ideas  from research labs to the real world- that focus started really, in my PhD, due to personal reasons. And the personal story there is that due to a death in the family, I realized what we do in the research labs could literally save lives. Ironically, I was working on the exact same thing that could have saved a life, which was on early cancer detection. But at that time, we lacked the manufacturing technologies to really bring that solution from a lab bench to someone who could use that. And that pushed me in the direction of needing to make sure there are advanced manufacturing technologies that can be used to make products that people on the street can go and buy in a store. It's not just confined to one lab.

So the goal is to scale up and bring these manufacturing technologies to the market?

Yes- to the market and to the end user, and making an impact on people, not just focusing on understanding the basic science, but actually using it for applications.

What research projects are you going to be working on here?

 I'm excited about two research projects that I'm working on right now. The theme of both of these projects is predicting pattern formation in manufacturing processes. The first project is on nanoscale additive manufacturing, which involves making complex 3D structures with nanoscale features. We recently scaled up the process and are doing pretty good in terms of throughput and resolution . But we are not doing very good in terms of predictive capability. If I want a specific structure, I don't necessarily know how to tune the parameters to get there. I want to understand this process better to a point where I can predict this process and tell you what the inputs should be for whatever your desired structure is.

The topic of the second project is self-organization or self-assembly, which could be as simple as the wrinkling of clothes, but finely tuned to generate features on the micro and nano scale. And the problem there is that it's very easy to make these structures, but very difficult to predict the inputs that would generate your desired structure. Again, similar themes, but then the physics of the two processes are fundamentally different. The goal here is to answer how do we generate and analyze new process knowledge about self-organization to a point where you can predict the inputs that give you a particular structure.

How many students do you plan to have working in your lab?

The guiding principle here is that I don’t want to treat my graduate students as an extra pair of hands. They're really going to be encouraged to take ownership of their own projects. So, the number of students will scale with the number of projects that I have. At the same time, I need to make sure that I can give them individual attention and guidance and mentoring. So that puts an upper bound on the number of students that I can have. At this time I do not know the right number for me, but I have started with one graduate student in my group.

What is the biggest challenge of being a new Professor on campus?

The closest analogy I can come up with is that being a new professor on campus is like trying to jump on to a moving train. That has been the biggest challenge. Academia is an enterprise where professors do a lot of different things. On the same day, you could be a teacher, you could be a mentor to your graduate students, could be a fundraiser, or just presenting to the external world what you're doing, and doing all of these things on a daily basis is not something that I've done in the past. I've done all of these things, but not necessarily on the same day or even in the same week. So that's a challenge. Our senior faculty seem to be pretty good at this, so I'm trying to ramp up to that. They have given me something to aspire to. 

The 2019 award-winning HBO miniseries Chernobyl tells the story of the 1986 nuclear plant explosion which took the lives of 28 workers and inflicted radiation symptoms on approximately 3.5 million people. Nuclear energy is a powerful proposition that can create clean energy for our planet, but disasters at plants such as Chernobyl and Fukushima and the threat of nuclear warfare keep the public wary.

Associate professor Anna Erickson was the featured guest on this month's episode of Dean Steve McLaughlin's Uncommon Engineer podcast to talk about these challenges. During the conversation she talked about Chernobyl, nuclear power and security, safety, detection, non-proliferation, and her new consortium tasked with helping to develop the next generation of nuclear safety experts. 

Download the audio file or listen via YouTube below.

On Wednesday, September 11, 2019, a breakfast was held to honor prestigious fellowship award winners from within the College of Engineering. Woodruff School students from the mechanical engineering, nuclear engineering, and bioengineering programs received a total of 49 fellowships awarded this year. The awards, students, and their advisors are:
 

Test subject who has flexible wireless electronics conformed to the back of the neck, with dry hair electrodes under a fabric headband and a membrane electrode on the mastoid, connected with thin-film cables. (Courtesy of Woon-Hong Yeo)

Combining new classes of nanomembrane electrodes with flexible electronics and a deep learning algorithm could help disabled people wirelessly control an electric wheelchair, interact with a computer or operate a small robotic vehicle without donning a bulky hair-electrode cap or contending with wires.

By providing a fully portable, wireless brain-machine interface (BMI), the wearable system could offer an improvement over conventional electroencephalography (EEG) for measuring signals from visually evoked potentials in the human brain. The system’s ability to measure EEG signals for BMI has been evaluated with six human subjects, but has not been studied with disabled individuals.

The project, conducted by researchers from the Georgia Institute of Technology, University of Kent and Wichita State University, was reported on September 11 in the journal Nature Machine Intelligence.

“This work reports fundamental strategies to design an ergonomic, portable EEG system for a broad range of assistive devices, smart home systems and neuro-gaming interfaces,” said Woon-Hong Yeo, an assistant professor in Georgia Tech’s George W. Woodruff School of Mechanical Engineering and Wallace H. Coulter Department of Biomedical Engineering. “The primary innovation is in the development of a fully integrated package of high-resolution EEG monitoring systems and circuits within a miniaturized skin-conformal system.”

BMI is an essential part of rehabilitation technology that allows those with amyotrophic lateral sclerosis (ALS), chronic stroke or other severe motor disabilities to control prosthetic systems. Gathering brain signals known as steady-state virtually evoked potentials (SSVEP) now requires use of an electrode-studded hair cap that uses wet electrodes, adhesives and wires to connect with computer equipment that interprets the signals.

Yeo and his collaborators are taking advantage of a new class of flexible, wireless sensors and electronics that can be easily applied to the skin. The system includes three primary components: highly flexible, hair-mounted electrodes that make direct contact with the scalp through hair; an ultrathin nanomembrane electrode; and soft, flexible circuity with a Bluetooth telemetry unit. The recorded EEG data from the brain is processed in the flexible circuitry, then wirelessly delivered to a tablet computer via Bluetooth from up to 15 meters away.

Beyond the sensing requirements, detecting and analyzing SSVEP signals have been challenging because of the low signal amplitude, which is in the range of tens of micro-volts, similar to electrical noise in the body. Researchers also must deal with variation in human brains. Yet accurately measuring the signals is essential to determining what the user wants the system to do.

To address those challenges, the research team turned to deep learning neural network algorithms running on the flexible circuitry.

“Deep learning methods, commonly used to classify pictures of everyday things such as cats and dogs, are used to analyze the EEG signals,” said Chee Siang (Jim) Ang, senior lecturer in Multimedia/Digital Systems at the University of Kent. “Like pictures of a dog which can have a lot of variations, EEG signals have the same challenge of high variability. Deep learning methods have proven to work well with pictures, and we show that they work very well with EEG signals as well.”

In addition, the researchers used deep learning models to identify which electrodes are the most useful for gathering information to classify EEG signals. “We found that the model is able to identify the relevant locations in the brain for BMI, which is in agreement with human experts,” Ang added. “This reduces the number of sensors we need, cutting cost and improving portability.”

The system uses three elastomeric scalp electrodes held onto the head with a fabric band, ultrathin wireless electronics conformed to the neck, and a skin-like printed electrode placed on the skin below an ear. The dry soft electrodes adhere to the skin and do not use adhesive or gel. Along with ease of use, the system could reduce noise and interference and provide higher data transmission rates compared to existing systems.

The system was evaluated with six human subjects. The deep learning algorithm with real-time data classification could control an electric wheelchair and a small robotic vehicle. The signals could also be used to control a display system without using a keyboard, joystick or other controller, Yeo said.

“Typical EEG systems must cover the majority of the scalp to get signals, but potential users may be sensitive about wearing them,” Yeo added. “This miniaturized, wearable soft device is fully integrated and designed to be comfortable for long-term use.”

Next steps will include improving the electrodes and making the system more useful for motor-impaired individuals.

“Future study would focus on investigation of fully elastomeric, wireless self-adhesive electrodes that can be mounted on the hairy scalp without any support from headgear, along with further miniaturization of the electronics to incorporate more electrodes for use with other studies,” Yeo said. “The EEG system can also be reconfigured to monitor motor-evoked potentials or motor imagination for motor-impaired subjects, which will be further studied as a future work on therapeutic applications.”

Long-term, the system may have potential for other applications where simpler EEG monitoring would be helpful, such as in sleep studies done by Audrey Duarte, an associate professor in Georgia Tech’s School of Psychology.

“This EEG monitoring system has the potential to finally allow scientists to monitor human neural activity in a relatively unobtrusive way as subjects go about their lives,” she said. “For example, Dr. Yeo and I are currently using a similar system to monitor neural activity while people sleep in the comfort of their own homes, rather than the lab with bulky, rigid, uncomfortable equipment, as is customarily done. Measuring sleep-related neural activity with an imperceptible system may allow us to identify new, non-invasive biomarkers of Alzheimer's-related neural pathology predictive of dementia.”

In addition to those already mentioned, the research team included Musa Mahmood, Yun-Soung Kim, Saswat Mishra, and Robert Herbert from Georgia Tech; Deogratias Mzurikwao from the University of Kent; and Yongkuk Lee from Wichita State University.
 

This research was supported by a grant from the Fundamental Research Program (project PNK5061) of Korea Institute of Materials Science, funding by the Nano-Material Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (no. 2016M3A7B4900044), and support from the Institute for Electronics and Nanotechnology, a member of the National Nanotechnology Coordinated Infrastructure, which is supported by the National Science Foundation (grant ECCS-1542174).

CITATION: Musa Mahmood, et al., “Fully portable and wireless universal brain-machine interfaces enabled by flexible scalp electronics and deep learning algorithm.” (Nature Machine Intelligence, 1, 412-422, 2019). https://doi.org/10.1038/s42256-019-0091-7

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA
Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu).
Writer: John Toon

Aaron Young and Greg Sawicki in their lab space

Aaron Young and Greg Sawicki approach robotics from different angles, but they share the same goal- they want to develop wearable robotic devices that can help people maintain or increase their mobility. Together their labs in Georgia Tech’s George W. Woodruff School of Mechanical Engineering are laying the groundwork for a range of devices that could help children, stroke patients, the elderly, and amputees gain more independence and move with greater comfort and ease.

“My goal in the lab is to make sure that people can do what they want to do as long as they want to,” says Sawicki. “That might be a big challenge, but it’s a straightforward guiding principle- make people happy by allowing them to keep moving.”

Aaron Young approaches these challenges as a controls specialist focused on how the human body can interface with a robotic device. The human body processes countless signals every time a person moves, so he is focused on translating those neural signals into commands for a range of exoskeletons and prostheses that can perform as seamlessly as if they were a part of the body.

"I was always really interested on the control side of things, understanding how to take biological signals or neural signals from the human and use that to control robotic devices," says Young. "That was really my focus as a graduate student at Northwestern and what got me into this field of human augmentation and assistive restoration with prostheses and exoskeletons."

"I'm trying to understand how to unite the human user with their assistive robotic technology. A major part of that focus is on something that we call intent recognition which is really trying to understand what a person is trying to do and then use different artificial intelligence and machine learning techniques to take sensor information from the human and the robot and use that to appropriately give assistance."

With Young focused on the human-machine interface and control systems, Sawicki looks at the neuromechanics and energetics of human locomotion and how movement can be augmented with wearable technology. An avid runner and biker, he has always been fascinated with the efficiency of movement, and he is focused on finding ways to restore that efficiency to people who are lacking in it. If that can be done, they could live more active lifestyles over a longer period of time. Sawicki also looks at the impact wearable devices have on the body itself, whether that’s in terms of energy cost or how something attached to a person’s ankle might impact their knee or hip.

As he puts it, “When I got involved with wearable robotics I saw that we're attaching something to the outside of a very complex structure- the human body. And if you don't think at all about what's happening under the skin you're not going to solve the problem. You have to understand muscles and how they work to build good robots.”

And make no mistake- Sawicki and Young are building good robots, even though they might not be what comes to mind when most people think of robots.

Their devices are wearable, and have varying degrees of electronics and coding built into them. Though the technology behind them varies, they all share a common goal- to make it easier for people to be mobile. Here are some of the projects they are working on.

Knee Exoskeleton

Young is developing a lightweight low-profile knee exoskeleton for children with cerebral palsy that helps them counteract a condition known as crouch gait. The condition is caused by perpetually tense muscles and tendons, and the knee exoskeleton is being designed to act as a mechanical intervention that helps retrain the tendons, muscles, and the brain by reinforcing a more biomechanically efficient gait. Right now treatment options include physical therapy and surgical procedures and the hope is that this device could help patients avoid surgical intervention by aiding therapist during physical therapy. The project is being sponsored by Children’s Healthcare of Atlanta and the Atlantic Pediatric Device Consortium. 

“The nice thing about mechanically assistive robotics is a therapist could use it to create new training paradigms for children with cerebral palsy to ultimately improve their long term clinical outcomes using these devices,” said Young.

Robotic Prosthesis

Clinician Kinsey Herrin and Aaron Young with a patient wearing the robotic lower limb.

The robotic leg prosthesis is a two joint device, with powered knee and ankle joints. Most prosthetic legs are passive, using no power, and they are optimized for performing one task- walking on level surfaces. Young wanted to see what would happen if a device was designed to be able to adapt to a user’s needs, whether that’s climbing stairs, walking on a ramp, or moving around on flat ground. A major part of that is trying to understand when someone wants to transition between these different tasks and to be able to give the appropriate assistance at the right time. The goal is for it to be natural, automatic, and intuitive to the user with no need to hit buttons or switches during transitions. Instead the objective is to develop an intent recognition system that employs a variety of machine learning and artificial intelligence techniques to interpret sensor information and anticipate what the user is going to do next.

Autonomous Hip Exoskeleton

Designed primarily for stroke patients who have lost a degree of their mobility, the powered hip exoskeleton offers assistance in hip flexion and extension. Like the leg prosthesis, the amount of power need can fluctuate based on what the user is trying to do and whether they want to move faster or slower, up stairs, or down ramps. The device could be used to help users move around more comfortably and independently in their homes, or as a therapy device in a rehabilitation environment.

“The user may be able to walk but, they are not walking as well as they were before their injury,” said Young. “The ability to restore full community ambulation is really valuable for improving independence, quality of life, and overall happiness of any individual. Our focus is on trying to understand how to optimize the control interface so that it feels natural so the patient enjoys using it and is able to more easily get around through this augmented assistance.“

This project is being funded through the National Science Foundation (NSF) National Robotics Initiative and a National Institutes of Health (NIH) National Center for Medical Rehabilitation Research Young Investigator Award recently awarded to Young.

Ankle Exoskeleton

While the first three projects are focused on relatively narrow segments of the population with hopes of them being expanded down the road, an ankle exoskeleton is being developed for the aging population.

As Greg Sawicki describes it, the Achilles tendon is the powerhouse spring in the ankle, and as people age that spring loses its stiffness and subsequently loses its ability to store and return energy during a step. That results in more effort and energy being used by muscles to do the same amount of work that was once accomplished with vibrant tendon recoil. The ankle exoskeleton Sawicki’s team is developing uses a spring and clutch system to restore that bounce and compensate for the body’s natural decline, allowing people who have perhaps transitioned from the workforce into retirement to maintain active and independent lifestyles.

Although each of these projects is different there’s an underlying principle behind each of them that focuses on efficient mobility. The human body is an incredibly complicated machine, and as one part performs less than optimally other parts compensate, often to their detriment. For example, amputees often have higher rates of osteoarthritis and osteoporosis. There’s also the matter of fuel efficiency. Sawicki compares the body to a car- as parts wear out the efficiency of the vehicle declines, requiring more fuel to go the same distance. As an Achilles tendon loses its snap, leg muscles at the hip or knee must work harder, consuming more energy. By taking preventative steps through human augmentation with external, wearable robotics Sawicki and Young think they can help stave of decline in other parts of the body and extend the active and productive years of an aging population.

To accomplish all of this Young and Sawicki have a state of the art workspace staffed with interdisciplinary teams of postdocs, graduate students, and undergraduate researchers working together through Young’s Exoskeleton and Prosthetic Intelligent Controls (EPIC) lab and Sawicki’s Physiology of Wearable Robotics (PoWeR) lab to tackle the engineering, programming, and clinical challenges of restorativedevices. Truly interdisciplinary in nature, their research leverages the expertise and resources of a range of affiliations. Both Sawicki and Young belong to Georgia Tech’s Institute for Robotics and Intelligent Machines (IRIM) and are members of the Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory program faculty as well as members of Tech’s Parker H. Petit Institute for Bioengineering and Bioscience. Sawicki also has an appointment to Georgia Tech’s School of Biological Sciences. Those interdisciplinary connections allow the pair to attract students from across schools and majors.

Graduate students with the robotic lower limb they helped build and design using parts fabricated at Georgia Tech.

Mechanical engineers in the lab are focused on the design, fabrication, and refinement of the devices. Electrical and computer engineers develop the controls and wiring of the systems, and biomedical engineers look at how the devices interact with their users and help to understand the medical and biomechanical consequences of the projects. With a wide array of equipment that includes motion capture, muscle monitoring, and force plates throughout their labs that can capture, measure, and model the forces and consequences of every movement, team members can make sure that the exoskeletons and prostheses are functioning as intended and not adding extra strain to other parts of the body. The labs also include a team of computer scientists focused on developing the machine learning and artificial intelligence that serve as the foundation for the assistive technology.

On the clinical side, particularly with the lower limb prosthesis, the labs work with Georgia Tech’s master’s in prosthetics and orthotics (MSPO) program to find volunteers interested in helping to test the devices. The clinicians help to make sure the devices are properly aligned and safe to wear.

“The patients we recruit like being part of this,” says Kinsey Herrin, the clinical liaison and instructor in the Prosthetics and Orthotics program. “They like giving back and feeling like they're contributing to science. They enjoy the energy of it and being on the edge of something new.”

The collaboration between engineering, biology, computer science, and clinical medicine is one of the things that sold Aaron Young on coming to Georgia Tech as a faculty member.

“I think what is really unique about Georgia Tech is not only do you have the clinical resources of a large city with good supporting hospitals, but you have the really advanced research and engineering apparatus to partner with, especially on things like artificial intelligence, machine learning, and mechanical engineering. We have the capability to do design creation and prototyping on site and that really helps to set us apart from other places. As an incoming faculty member it's a place that's really attractive to do this kind of research where we can build everything at home, do a lot of good testing and implement lots of very interesting things. We have an amazing robotics community here, and that makes it a lot easier to do the interdisciplinary research that we're really trying to do.”

“I think we have one of the best facilities in the world now,” adds Sawicki. “We’re definitely one of the best in the nation for assessing how robots interact with people all the way from under the skin up through the whole body. We’re especially focused in this space on being able to understand how these things work out in the real world environment. It’s time to cross that hurdle now.”
 

This research was supported by grants to G.S.S. from the National Robotics Initiative via the National Institute of Nursing Research of the National Institutes of Health (R01NR014756), the National Institute on Aging of the National Institutes of Health (R01AG058615) and the U.S. Army Natick Soldier Research, Development and Engineering Center (W911QY18C0140).

This research was supported by grants to A.J.Y. from the NSF National Robotics Initiative (Award #1830215), the Office of the Assistant Secretary of Defense for Health Affairs through the Orthotics and Prosthetics Outcomes Research Program Prosthetics Outcomes Research Award under Award No. W81XWH-17-1-0031, and an NIH New Investigator NCMRR award (Grant Number: 1R03HD097740-01).

The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies listed.

Student Testimonials

Dawit Lee
Mechanical engineering master’s student
Seoul, Korea
Lead on of the pediatric knee exoskeleton project

“The biggest challenge from the project is that we are dealing with humans. It's not a robotic autonomous device that operates by itself. Rather, we have to consider the human aspect that includes the interaction between the human and the robot, that the human itself is already a design system. We have to build a smart control system on the robotic side that can work well with the human. To maximize the effectiveness of wearable robotic devices.

The most rewarding part of this project is seeing is a new emerging technology that can assist our human living, helping us to move easier or longer. Hopefully in the future some of the children who have difficulty walking can benefit by using this type of robotic devices.”

Inseung Kang
Ph.D. student in mechanical engineering
South Korea
Hip exoskeleton for stroke patients

“During my undergrad at Georgia Tech I got a chance to work at a research lab that was making exoskeletons to help spinal cord injury patients. That's where I was exposed to wearable robotics and trying to help people with walking disabilities. That's what inspired to get a Ph.D. degree to continue that work.

From a technical standpoint the biggest obstacle for us is that we're focusing on autonomous solutions but all of these robotic devices require other technology to advance at the same time- like battery life. All of these technologies have to progress together for us to be able to make our product work in real world settings. That's the challenge of making these robots. Before I graduate I would like to see the stroke patients see the benefits of our research. That would be a dream come true.”

Krishan Bhakta
Ph.D. student in mechanical engineering
Santa Fe, New Mexico
Powered knee-ankle prosthetic device

“I've always enjoyed tinkering with things, so I wanted to find something that could apply engineering principles in a way that allowed me to help other people.

What I've enjoyed the most has been building the device. On this project I've been able to help out with a lot of the development behind what it is actually necessary to get something to work in a manner that makes it easier for people to use it. This goes beyond just the type of technology, but in general, what it takes to develop something that can be used easily by someone.

I had no idea Georgia Tech had facilities like this. This was the last school I applied to but the first to get back to me. That really impressed me. I didn't know anything about the resources they had available here. It's been amazing. I have no regrets about coming here, and that's always a good sign.”

Lindsey Trejo
PhD student in bioengineering
Lincoln, Nebraska
Ankle exoskeleton project for the elderly

“I've always been interested in rehabilitation and helping others, and I had done an engineering degree as an undergrad, but didn't have as much of an understanding of how the body works, and how to apply it to the body. That was in biological systems engineering. From my master's I studied kinesiology- what is the body doing and how is it supposed to be moving. My doctorate really allowed me to combine the two.

This is a really great lab- not many labs combine the kinesiology aspect with engineering. A lot of labs are just kinesiology or just engineering. This is a great one for both.

The resources here are great. I've been in other labs, and what we have here is amazing. We have a lot of cameras, and a lot of really new equipment. With this equipment we can get high quality data and really test our hypotheses.”

Bailey McLain
Biomedical engineering undergraduate
River Falls, Wisconsin
Knee exoskeleton

“I've always been interested in the prosthetic, orthotic, exoskeleton area. That's why I pursued my degree in BME. But this project specifically, being able to do a biomechanical analysis, and actually see how this exoskeleton could help people, and the potential for helping children who have trouble walking- it really hit home. I thought it was a great project to work on.

I love doing human experiments and being able to analyze the data. Going from the data collection to seeing what the results show is probably my favorite thing about it. Most of my fellow undergrads think it's really cool that I get to do research like this. I’m a Petit Scholar so I’ll be working on this for three more years.”

Using a new time-based method to control light from an ultrafast laser, researchers have developed a nanoscale 3D printing technique that can fabricate tiny structures a thousand times faster than conventional two-photon lithography (TPL) techniques, without sacrificing resolution.

Despite the high throughput, the new parallelized technique — known as femtosecond projection TPL (FP-TPL) — produces depth resolution of 175 nanometers, which is better than established methods and can fabricate structures with 90-degree overhangs that can’t currently be made. The technique could lead to manufacturing-scale production of bioscaffolds, flexible electronics, electrochemical interfaces, micro-optics, mechanical and optical metamaterials, and other functional micro- and nanostructures.

The work, reported Oct. 3 in the journal Science, was done by researchers from Lawrence Livermore National Laboratory (LLNL) and The Chinese University of Hong Kong. Sourabh Saha, the paper’s lead and corresponding author, is now an assistant professor in the George W. Woodruff School of Mechanical Engineering at the Georgia Institute of Technology.

Existing nanoscale additive manufacturing techniques use a single spot of high-intensity light — typically around 700 to 800 nanometers in diameter — to convert photopolymer materials from liquids to solids. Because the point must scan through the entire structure being fabricated, the existing TPL technique can require many hours to produce complex 3D structures, which limits its ability to be scaled up for practical applications.

“Instead of using a single point of light, we project a million points simultaneously,” said Saha. “This scales up the process dramatically because instead of working with a single point that has to be scanned to create the structure, we can use an entire plane of projected light. Instead of focusing a single point, we have an entire focused plane that can be patterned into arbitrary structures.”

To create a million points, the researchers use a digital mask similar to those used in projectors to create images and videos. In this case, the mask controls a femtosecond laser to create the desired light pattern in the precursor liquid polymer material. The high-intensity light causes a polymerization reaction that turns the liquid to solid, where desired, to create 3D structures.

Image shows a printed micropillar forest submerged in the photopolymer resist prior to development. The forest contains 900 micropillars over a 7 mm × 7 mm area and was printed in less than 90 minutes as compared to more than a day of printing with serial techniques. (Credit: Vu Nguyen and Sourabh Saha)
 

Each layer of the fabricated structure is formed by a 35-femtosecond burst of high-intensity light. The projector and mask are then used to create layer after layer until the entire structure is produced. The liquid polymer is then removed, leaving behind the solid. The FP-TPL technique allows the researchers to produce in eight minutes a structure that would take several hours to produce using earlier processes.

“The parallel two-photon system that has been developed is a breakthrough in nanoscale printing that will enable the remarkable performance in materials and structures at this size scale to be realized in usable components,” said LLNL’s Center for Engineered Materials and Manufacturing Director Chris Spadaccini.

Unlike consumer 3D printing that uses particles sprayed onto a surface, the new technique goes deep into the liquid precursor, allowing the fabrication of structures that could not be produced with surface fabrication alone. For instance, the technique can produce what Saha calls an “impossible bridge” with 90-degree overhangs and with more than a 1,000:1 aspect ratio of length to feature size. “We can project the light to any depth that we want in the material, so we can make suspended 3D structures,” he said.

Overhanging 3D structures printed by stitching multiple 2D projections, demonstrating the ability to print depth-resolved features. The bridge structure, with 90-degree overhangs, is challenging to print using serial scanning TPL techniques. (Credit: Vu Nguyen and Sourabh Saha)
 

The researchers have printed suspended structures a millimeter long between bases that are smaller than 100 microns by 100 microns. The structure doesn’t collapse while being fabricated because the liquid and solid are about the same density — and the production happens so quickly that the liquid doesn’t have time to be disturbed.

Beyond bridges, the researchers made a variety of structures chosen to demonstrate the technique, including micro-pillars, cuboids, log-piles, wires and spirals. The researchers used conventional polymer precursors, but Saha believes the technique would also work for metals and ceramics that can be generated from precursor polymers.

“The real application for this would be in industrial-scale production of small devices that may be integrated into larger products, such as components in smartphones,” he said. “The next step is to demonstrate that we can print with other materials to expand the material palette.”

Research groups have been working for years to accelerate the two-photon lithography process used to produce nanoscale 3D structures. The success of this group came from adopting a different way of focusing the light, using its time-domain properties, which allowed production of very thin light sheets capable of high resolution — and tiny features.

Use of the femtosecond laser allowed the research team to maintain enough light intensity to trigger the two-photon process polymerization while keeping the point sizes thin. In the FP-TPL technique, the femtosecond pulses are stretched and compressed as they pass through the optical system to implement temporal focusing. The process, which can generate 3D features smaller than the diffraction-limited, focused light spot, requires that two photons hit the liquid precursor molecules simultaneously.

“Traditionally, there are tradeoffs between speed and resolution,” Saha said. “If you want a faster process, you would lose resolution. We have broken this engineering tradeoff, allowing us to print a thousand times faster with the smallest of features.”

At Georgia Tech, Saha intends to continue advancing the work with new materials and further scale-up of the process.

“So far, we have shown that we can do pretty well on speed and resolution,” he said. “The next questions will be how well we can predict the features and how well we can control the quality over large scales. That will require more work to understand the process itself.”

CITATION: Sourabh K. Saha, Dien Wang, Vu H. Nguyen, Yina Chang, James S. Oakdale, Shih-Chi Chen, “Scalable submicrometer additive manufacturing.” (Science 2019). http://dx.doi.org/10.1126/science.aax8760

Aerodyme began as a Capstone Senior design project in the spring of 2019. Now the startup is taking preorders for it's innovative device designed to reduce drag on tractor trailers by more than 10%, improving fuel economy. ME graduate Jayce Delker and current student Tyler Boone talked to Dean Steve McLaughlin on his Uncommon Engineer podcast about their love for engineering, their experience as aspiring entrepreneurs, and the steps they've gone through to get the company off the ground, including participating in CREATE-X. 

Watch the Video:

Download the audio file

Read the Transcript