Howard, Ayanna

views updated

Ayanna Howard

1972—

Robotics engineer

Ayanna Howard was a child with science-fiction fantasies. When she encountered the 1970s' television series Bionic Woman—in which a severely injured woman was outfitted with robotic limbs and thus endowed with superhuman powers—Howard was less interested in becoming a super heroine then she was in understanding how the show's fictional scientists managed to create the robotic limbs that gave the bionic woman her powers. Howard was in elementary school at the time, but this early inspiration motivated her through years of hard work and meticulous research to become one of the twenty-first century's leading robotics engineers, helping to design the next generation of robots and making yesterday's science-fiction fantasies a modern reality.

Howard was born in 1972 in Providence, Rhode Island, but spent most of her childhood in California. Her family relocated to Inglewood, California, when Howard was between two and three years old and later to the Altadena/Pasadena area, where Howard's mother, Johnetta, and father, Eric, owned and operated a small company that manufactured parts for railroad signal stations. Howard's parents fostered and encouraged an early interest in machines and engineering.

Howard attended public schools in Pasadena, where she excelled in math and physics. Though she had an obvious aptitude for working with computers, machines, and electronics, she also recalls wanting to take a different path from that of her father. "My father was an engineer," Howard told Contemporary Black Biography (CBB), "so I knew I didn't want to do that." From age eleven, Howard believed she knew the course of her career: she would attend medical school, become a doctor, and specialize in limb replacement. However, her biology class during her freshman year in high school changed her mind about the future. "After taking biology, I realized that what I loved was math and physics and that what interested me about robotics was the mechanics and control, not the human physiology," Howard recalled in a 2002 interview with Imagiverse. org. By the time she was ready for college, she had decided, like her father, that her talents and interests were more geared toward engineering.

Returned to Providence

When Howard was applying to universities, her parents told her that they would only pay for her education if she left California, both for the experience of living in another state and because, as Howard told Stephen Cass of IEEE Spectrum Online, "They felt California was too laid back!"

Howard chose to attend Brown University in Providence, Rhode Island, because, as she told CBB, "Brown didn't use the grade system. I had always been competing for grades just to get into college and so, when I got to college, I didn't want to do that anymore." Freed from the constant concern of maintaining a grade point average, Howard allowed herself the freedom to explore Brown's liberal arts classes, taking dance and literature courses to supplement her core curriculum. Though Brown was a liberal university, the school offered a number of classes in technology and computing, and Howard decided to pursue a bachelor of science in computer engineering.

During her summer break in 1990, Howard returned to Pasadena, where she was accepted for an internship at the National Aeronautics and Space Association's (NASA) Jet Propulsion Laboratory (JPL). Howard enjoyed the JPL program and returned in 1991 and 1992 to help JPL engineers design a database. While at the JPL, Howard learned how to function in the applied research community and was exposed to new ideas and research interests in disparate fields. During her senior year internship, she began working with the JPL robotics group.

Entered the Robotics Field

Howard returned to Pasadena after graduation and entered the University of Southern California (USC) at Los Angeles for graduate studies. In 1994, during her second semester at USC, Howard was offered a position in the JPL's computer science department to work on software applications. Her work at NASA complemented her studies as she worked toward a master's in electrical engineering. While at USC, Howard became fascinated with the concept of artificial intelligence (AI). By 1997, when she began her PhD program in USC's engineering department, Howard was working on developing software for AI applications.

In 1999, the same year she received her PhD, Howard joined the JPL's Telerobotics Research and Information Group. Though her initial interest was in building the mechanical components used in robotics, Howard's experience with AI and programming led her to the Dynamic Programming Group, where she was introduced to nonlinear algorithms and neural network programming.

While working as a junior robotics researcher, Howard began publishing papers on robot "thinking" and the use of software systems in directing navigational systems. From 1993 to 2005, Howard's research was featured in over sixty articles. In 2001, she was honored by the JPL with the Lew Allen Award for Excellence in Research, which recognizes success in research and effectiveness as a teacher and leader in the field. In 2002, Howard was named as a senior robotics researcher.

Developed a Variety of Technologies

While working with the JPL, Howard's research was used for a number of different projects, of which space exploration and military applications absorbed most of her time. From 1998 to 2005, she spent some of her time working on software for military projects. Though Howard is now able to discuss some of the military projects she worked on at the JPL, some of her research applications are still classified as military secrets. One of Howard's declassified military projects was an application that uses visual recognition software to help identify dangerous elements on the battlefield.

At a Glance …

Born Ayanna MacCalla in 1972 in Providence, RI; daughter of Eric and Johnetta McCallum; married 2002 (divorced); child: Zyare Howard. Education: Brown University, BS computer engineering, 1993; University of Southern California, MS electrical engineering, 1997, PhD electrical engineering, 1999; Claremont University, MBA, 2005.

Career: National Aeronautics and Space Administration, Jet Propulsion Laboratory (JPL), computer information internships, 1993-97, information systems engineer, 1997-99, senior robotics researcher, 1999-2005; Georgia Tech University, associate professor, 2006—.

Awards: JPL, Lew Allen Award for Excellence in Research, 2001; Massachusetts Institute of Technology's Technology Review Top 100 Young Innovator of the Year, 2003; LA Council of Engineers and Scientists, Engineer of the Year Award, 2004; Allstate Insurance Distinguished Honoree for Achievement in Science, 2004; Young Global Leaders Award, 2006.

Addresses: Office—Van Leer Electrical Engineering Building, 777 Atlantic Dr. NW, Atlanta, GA 30332-0250.

"We were working with [Principal Component Analysis]," Howard told CBB, "a tool that picks out dominant features of an object or potential enemy and then classifies them as, for example, a jet, helicopter or missile." To create an automated program that could identify potential threats, researchers first needed to train their neural network by showing it thousands of examples of different objects that fall into target categories. Using a system similar to the way humans identify and classify objects, the computer learns to identify key features—such as the shape of the object's body or the presence or absence of wings and/or rotor blades—to classify the object into one of several predetermined classes. "We eventually achieved about a 96 percent recognition rate," Howard said. Visual recognition software can be used to identify the locations of enemy equipment and troops and from there scientists can use additional computational systems to infer information about enemy strategy.

The majority of Howard's time at the JPL was focused on developing technology for robotic exploration of extreme environments, such as those found on other planets. Howard was a leader of the Safe Rover Navigation Task, a project aimed at building robots that can navigate independently on the Martian landscape. The rocky terrain of Mars makes navigation difficult, and Howard and her colleagues realized that, to navigate successfully, robot rovers will need the capability to evaluate the terrain visually, using a system superficially similar to that of a human explorer.

Howard explained in an interview with Time's Dan Cray, "People always look for the straightest, clearest path, so that's what we map to the robot." The goal is to create robots that can function with relative autonomy. Howard's neural network programming enables the robots to integrate new data into an evolving behavioral scheme, thereby allowing adaptation. Though the process is different from that used by humans, Howard's robots are "thinking" machines. "They recognize things like rock distribution, slope, and the presence or absence of cliffs and depressions," Howard told CBB. "It's a fuzzy logic engine that says, for instance, ‘If rock distribution is high but the slope is low and there are no cliffs, then the terrain is navigable.’"

The ultimate goal is to free the operators from having to direct the robots' every movement, which is the system used for the current Mars rovers Spirit and Opportunity. In an interview with IEEE Spectrum Online, Howard said, "I want to plop a rover on Mars and have it call back when it finds interesting science." In Howard's estimation, it will eventually be possible to design robots that are capable of mineralogical analysis while moving through the environment. The rover will investigate rocks and samples that appear interesting, and will then use its initial analysis to determine whether further investigation is warranted.

Howard's designs will feature prominently in NASA's planned exploration of Mars, set to begin in December of 2009 and land on the Martian surface in 2010. The project will be part of NASA's Mars Science Laboratory, a moving laboratory that will collect geologic samples and eventually send samples back to Earth during future missions. Howard's designs for neural networks and automated navigation are at the forefront of robotics research and have applications beyond space exploration. In 2003, Howard was named as one of the world's top innovators by the Massachusetts Institute of Technology's Technology Review in recognition for her pioneering work in AI.

Expanded Focus to Community and Academia

As Howard gained recognition as a leader in her field, she began taking a role as a community leader, helping to disseminate information about science and engineering to young women in the Altadena community. Howard was a frequent guest at Southern California schools, where she gave speeches about employment opportunities in engineering. "So many of our girls get persuaded to pursue other career choices at a young age and become discouraged with math and science," Howard said in an interview with Imagiverse.org. "By the time robotics is thought of as a career choice, they have already been left behind in the game." From 2001 to 2003, Howard started a mentoring program for young women in the Altadena area, focusing on encouraging achievement in science and mathematics.

While working for the JLA, Howard began pursuing an master in business administration (MBA) in hopes of eventually transitioning into the administrative aspects of the research community. Howard felt that her MBA training was an immediate benefit because it helped her to learn skills that were not common among researchers, such as how to deal with the public or handle relational issues among employees and associates. "I have a bigger vision in terms of what I want to do when I'm 50," she said to Cass. "It's about being strategic in what you want to do."

In 2005, Howard left the JPL for an associate professor position at Georgia Tech University. The decision to leave the JPL was partially based on a massive restructuring at NASA, after which Howard found that the organization was no longer able to provide the compensation and research environment she desired. However, at Georgia Tech she found an environment conducive to research and was able to remain involved in helping young people develop an interest in technology.

After joining the university's Systems and Controls Group, she founded the Human Automation Systems Laboratory. One of Howard's primary goals is to work on technological innovations that help the field of robotics transition from strictly scientific applications to more general industrial functions. Howard believes that industrial robotics will grow rapidly in the twenty-first century and that AI research will play a major role in the future of robotics application.

Beginning in 2006, Howard spent some of her time at Georgia Tech working on social assistance applications. One of her goals is to develop neural networks that can learn from human demonstration. Howard wants to design a simple interface that will allow users with no experience in technology or robotics to train domestic robots. "Say you want a robot to help clean up the children's toys in the morning or feed your grandmother breakfast," Howard said to CBB. "We're now at a stage where simple robot technology is available, like the Roomba [a robotic vacuum cleaner that navigates a home while cleaning carpets]. Now we have to go beyond that." Howard is certain that assistive technology will be the most important field of robotics in the coming years.

Besides working with assistance technology, in 2007 Howard was also collaborating with NASA researchers on a project that has immediate applications for research on global warming. The project has allowed Howard and her colleagues to work with a new set of variables using groups of robots to navigate inhospitable areas, such as the polar ice caps, and cooperate to retrieve information about global warming. After reaching a target location, the probes will use visual cues to determine the best locations to look for data. By sharing information and comparing data, a group of robots will have a better chance of maximizing time spent looking for fruitful data collection areas. In the future, Howard hopes that robots of this kind will retrieve data that will be impossible or impractical for humans to obtain and will free operators to work on data analysis rather than on the mechanics of data collection.

Besides contributing to the development of robotics science, Howard hopes to continue working in the community. After obtaining a grant from the National Science Foundation in 2007, Howard began working on a program to create video games aimed at underrepresented students, such as women and minorities, to create interest and understanding in technology careers. The games will use simulations, such as space exploration and tracking fugitives as a law enforcement officer, to teach students about data retrieval and analysis. "They will be doing data mining," she said to CBB, "but they don't know it. If you have a bunch of data, how do you find something interesting and relevant?"

As of 2007, Howard was considered one of the nation's leading experts in robotics and AI. Working on pioneering projects in a number of fields, Howard has also become a leader in helping engender interest in science for the next generation of students. Her own experience as an African-American woman in a typically white male field helps her to understand the struggles that many young women endure. When Imagiverse.org asked her what advice she would give to students interested in engineering, she replied, "Don't let anybody persuade you to give up. Know that some things may be difficult and others may cause you to struggle, but deal with it and keep trying."

Selected writings

Nonfiction

(With Curtis Padgett) "A Generalized Approach to Real-Time Pattern Recognition in Sensed Data," Pattern Recognition, Vol. 32, 1999, pp. 2069-2071.

(With George A. Bekey) "Intelligent Learning for Deformable Object Manipulation," Autonomous Robots, Vol. 9, 2000, 51-58.

(With Homayoun Seraji and Edward Tunstel) "Safe Navigation on Hazardous Terrain," International Conference on Robotics and Automation, 2001, pp. 3084-3091.

(With Seraji and Tunstel) "A Rule-Based Fuzzy Traversability Index for Mobile Robot Navigation," International Conference on Robotics and Automation, 2001, pp. 3067-3071.

(With Curtis Padgett) "An Adaptive Learning Methodology for Intelligent Object Detection in Novel Imagery Data," Neurocomputing, Vol. 51, 2003, pp. 1-11.