Hale:C.1 Source Acquisition

Topic

The topic which I chose for this assignment is gait transitions in legged robots. Anecdotally, I've seen animals transition from running to walking to climbing to an assortment of other gaits (generally) without falling, tripping, or stumbling. I want to replicate this behavior in legged robots, thus allowing them to change gaits as the terrain on which they are walking changes or as other environmental conditions change in order to better respond to the demands of their changing surroundings. I know that several different types of gaits have been tuned and implemented in legged robots and while these gaits are effective of their own, there would be great value in identifying a means of transitioning between these gaits so that a particular robot may change between existing gaits in real time.

The particular question that will guide me is: "What enables animals transition from one gait to another while locomoting? How can these transitions be made, where possible, in robots?"

C.1.0) Source

The paper from bioinspired robotics that I am choosing as my inspiration for this assignment is [1]. This paper was published in 2006, meaning that it fulfills the requirement that our source for this portion of the assignment be no more than 5 years old. I have chosen it because it explores several different types of gaits on two different robotic platforms, thus providing a diversity of investigations while coinciding very nicely with my overarching question.

C.1.1) Venue

This paper came from the IEEE International Conference on Robotics and Automation (ICRA). ICRA is an annual, international conference which is hosted by the IEEE's Robotics and Automation Society. According to the website for this organization:

The [Robotics and Automation] Society strives to advance innovation, education, and fundamental and applied research in Robotics and Automation.

It is clear that the research-centric nature of this conference is in line with what we seek to find here. Further, the society which hosts this conference is a subsidiary organization of the IEEE. The IEEE is one of the world's most respected engineering organizations and greatly contributes to the credentials of this conference. Based on the stated goals of the organization hosting ICRA and based on the fact that this organization is a part of the IEEE, this conference appears to be very credible and thus a good venue in which to find a basis for this assignment.

C.1.2) Authors

The second author for this paper (who is the Principal Investigator) is now listed as Adjunct Faculty at the Robotics Institute at Carnegie Mellon as well as Head Robot Scientist at Boston Dynamics. As Dr. Rizzi holds positions at both a well-respected research university and one of the leaders in industry in robotics research, he gives significant credibility to this paper. The first author of this paper was a graduate student studying under Dr. Rizzi at this time of this paper's publication. He is now a Postdoctoral Researcher at the University of Pennsylvania for Professor Dan Koditschek. Given Dr. Haynes position at a well-known research university and considering that he works directly with Professor Koditschek who has an extensive research career which includes several landmark contributions to robotics, Dr. Haynes's present appointment suggests that he, too, is a credible author.

Scopus lists an h-index of 12 for Dr. Rizzi. In general, the Scopus h-index values tend to be conservative, though 12 is a respectable number. Scopus presently tracks 64 documents authored or co-authored by Dr. Rizzi and lists 62 co-authors and 597 total citations over 19 years worth of publications. From this it is clear that Dr. Rizzi has had a sustained, extensive research career. With this and his aforementioned affiliations in mind, Dr. Rizzi appears to be a credible author and I see no reason to question his expertise. Scopus lists an h-index of 2 for Dr. Haynes. While this is a low h-index, it is important to note that Scopus only tracks 3 documents authored or co-authored by Dr. Haynes. Further, these documents are only from the years 2006-2009. In that span, Scopus shows that Dr. Haynes has worked with 11 co-authors on these 3 papers and that they have a total of 32 citations. While these figures are clearly lower than those for Dr. Rizzi (and are "low" in general), it seems that they are in large part due to the relative youth of Dr. Haynes's research career. Dr. Haynes's number of citations is reasonable as his publications have not had much time to be cited. Due to this lack of time, however, his number of citations is not the best measure of his reliability. In addition, the field of robotics as a whole tends to emphasize conferences more than journals, meaning that a journal-centric archive such as Scopus is also not the final arbiter of credibility. Notwithstanding Dr. Haynes's brief research career, his present Postdoctoral Research appointment bolsters his credibility; taken in conjunction with Dr. Rizzi's establish credibility, it seems that the team of these two authors is a credible source for engineering research and that this paper carries significant weight. Therefore, I find this paper to be an appropriate "seed" paper for this assignment.

C.1.3) Source Identification Methods

I e-mailed Doug once I felt that I had found good sources for each part of this assignment and he seemed to think that they were credible as well. While he confirmed that many robotics papers would be from conferences and that conferences' credentials are generally harder to verify than journals' credentials, he told me that the sponsoring organization can be used to gauge the value of a particular conference. In discussing my sources below, I have mentioned the sponsoring organization(s) of each conference paper to follow Doug's suggestion and use the reputation of sponsors to judge the reputation of each conference.

Doug's further suggestions related to methods of searching. I found several of my sources through Engineering Village and Doug mentioned that Engineering Village tracks many different controlled and uncontrolled keywords. While I used the uncontrolled keyword "gait transition" to help in my search, Doug brought up the possibility of using the "gait transition" keyword to narrow my search and then further narrow down those search results using controlled keywords such as "sensor," "actuator," or "power source" and similar terms that will help me explore a particular means of understanding or implementing gait transitions in robots. At present, I feel that my topic is of the appropriate scope and I don't see the need to further narrow my search. However, I do see the value of being able to use a single click to explore a single aspect of gait transitions and, assuming that I continue to use this topic in future assignments, will surely use Doug's suggestion as a means of finding relevant research.

Doug's final suggestion was to use the papers I had and find the research upon which they relied and find subsequent research which relied upon the papers I had chosen. It hadn't occurred to me to do this, though I do see the value in locating where in the chronology of the gait transition literature my sources lie. A brief discussion of predecessor and successor literature to my "seed paper" is presented below.

Past Related Research

According to Scopus, my main paper cites a total of 12 other research papers. Two of these papers deal exclusively with developing legged robotic platforms; they serve as the basic foundation for this work. Six of these papers deal with developing robot gaits. These serve as good precursor literature to the paper I have chosen to work from and lay an experimental basis for the work done within it. Some rely on the robotic platforms developed in the two papers mentioned above, meaning that there is a discernible progression of development present here. Three of the papers cited by this one are directly from biological literature with bolster the claim that this research has been (and should be) rooted in biology. The remaining paper is on robot juggling; while juggling is a decidedly different robot behavior, I believe that this paper is relevant based on its applications of dynamical systems to robotics. This somewhat eclectic set of citations represents that this paper was exploring a relatively new area of robotics research and building upon prior research in a way not yet done. The absence of citations of other papers dealing with gait transitions suggests that such literature did not yet exist (or was not yet extensive). Further, the inclusion of citations of work studying both robot juggling and stick insect locomotion represents building upon previous research in a new way, a good indicator of the "newness" of this research.

Subsequent Related Research

Scopus lists only a single publication which quotes this paper, which is a Ph.D. thesis from MIT published in 2008. The thesis title is "Metastable legged-robot locomotion," which suggests that gait transitions are only a part of the author's work and that gait transitions are only relevant to the extent that they promote stability. While there are undoubtedly more citations for this paper which are not tracked by Scopus, their scarcity in journals suggests that there have not been many subsequent studies into gait transitions. Thus, it appears that gait transitions are a worthy area to be explored in robotics, as it is still clearly an open problem.

C.1.4) High Quality Bioinspired Robotics Contribution

Having established that [1] is a credible, worthwhile contribution to robotics research literature, we may return to it to further discuss the problem being addressed here. In particular, the authors note that robots need gait transitions in order to make full use of their repertoire of behaviors:

Legged robots often utilize collections of gait patterns to locomote over a variety of surfaces. Each feedforward gait is generally tuned for a specific surface and set of operating conditions. To enable locomotion across a changing surface, a robot must be able to stably change between gaits while continuing to locomote.

This paper considers a small, though important, subset of gaits and focuses exclusively on transitions between them. Thus while it makes important strides in the area of legged locomotion, we see that robots must be able to transition effectively among all the gaits available to them (where transitions are possible) in order to be as effective in locomotion as animals.

C.1.5) General Robotics Literature

My First Source

My first of two papers from General Robotics Literature which presents research into gait transitions and further represents their validity is [2]
In particular, the introduction to this paper says that

Gait transitions are ubiquitous among legged animals and essential for robots.

which relates the robotics research in question to biology. Further, while this paper has explored several gait transitions in two robotic platforms, it readily acknowledges that

the more limited capabilities of legged robots [in comparison to animals] ensure for years to come that different maneuvers in different environments at different speeds under varied loading conditions will require the adoption of distinct locomotion patterns, along with necessitating the ability to transition between them safely and efficiently.

This makes clear that robots have not yet caught up to animals in their ability to transition between gaits in animals and provides further motivation for research.

The first author of this paper is Dr. Haynes, the same first author as the other paper. His is discussed above. Because he is an author of both of these papers, it is important to very closely examine the othor authors and venue of this paper to establish credibility outside Dr. Haynes's own.

Authors

First, we may look at the other two authors on this paper, the first of which is F. R. Cohen. There appears to be more than one profile for F. R. Cohen on both Google Scholar and Scopus, which complicates the process of exploring his credentials. However, the most complete profile of Dr. Cohen on Scopus shows that he has published papers from 1978 to 2010, meaning that he has had a 32-year research career. Further, he is presently a tenured Professor at the University of Rochester. However, the sum of the h-indices of F. R. Cohen's three Scopus profiles is only 6, which is a low number, especially considering the length of Dr. Cohen's career. Contrastingly, Google Scholar gives F. R. Cohen an h-index of 20 and 1978 total citations, which are much better numbers than those listed by Scopus and, while not extremely high, suggest that Dr. Cohen is a reasonably reputable source. While Google Scholar's h-indices are undoubtedly imperfect, Scopus tends to have very conservative numbers and does not account for conferences, meaning that Google Scholar's 20 value (likely) better captures Dr. Cohen's contributions to research than does Scopus's 6 value. Given the length of Dr. Cohen's career, his position as a Professor, and his (arguably) good h-index, he seems to be a credible author, though I do not believe that his credentials alone necessarily make this a credible source.

We may now examine the third author of this paper, Dr. Koditschek, who is this paper's principal investigator. Scopus gives Dr. Koditschek an h-index of 20 and 2068 total citations with publications spanning the years 1982 to 2009. Google Scholar gives Dr. Koditschek an h-index of 40 and 6303 total citations. As was stated previously, Scopus ignores conferences and robotics is a very conference-heavy field, thus leading me to give more weight to the Google Scholar figures. An h-index of 40 is a very good number, and 6303 citations suggest that Dr. Koditschek's research career has been both successful and influential. Further, the fact that he has published papers across a span of nearly 30 years lends credibility to his position as a researcher and indicates that he has had a sustained, successful career. While the first two authors of this paper have fairly good credentials, we see that this paper's Principal Investigator is very prominent in his field. Because Dr. Koditschek is Principal Investigator, I am inclined weight his credentials most in establishing the credibility of this paper and thus find that it is in fact a worthwhile contribution to robotics literature.

Note on Repeated Authorship

While Dr. Haynes appears on both papers, we need not worry about his repeated authorship as these papers differ in content, have different co-authors, and have both been peer reviewed. Further, these papers have different Principal Investigators, meaning that each project was lead by a different person. It does not appear that there is very much overlap between these two papers and I do not see any reason to be concerned about their content based upon any of the authors' credentials.

Venue

This paper was published in at the International Symposium of Robotics Research. The stated objective of this conference is:

The goal of the ISRR Symposia is to bring together active leading robotics researchers and pioneers from academia, government, and industry to assess and share their views and ideas about the state of the art in robotics, and to discuss promising new avenues for future research.

Based on their website, this conference appears to be held ever two years and is international. This conference has been held for over 20 years, which means that it has made early and continuous contributions to robotics. Further, the organization sponsoring ISRR is the International Foundation of Robotics Research, an organization which sponsors 5 total robotics symposia of different types. Given the age of the IFRR and the breadth which its symposia cover within robotics, it appears that the IFRR is a reputable organization and that ISRR is a worthwhile conference. The venue of this paper appears to lend credibility to this paper and bolster the notion that it is a meaningful contribution to robotics literature.

My Second Source

The second paper which I will choose to use from general robotics literature is [3]. The very first line of the abstract of this paper concisely identifies the importance of gait transitions by stating that the ability to traverse different terrains is essentially required for a robot to be relevant in the real world:

The ability to traverse a wide variety of terrains while walking is basically a requirement for performing useful tasks in our human centric world.

Thus we see that the ability to transition among gaits is very important for a robot that is to be used in diverse tasks and terrains as, based upon my two previous sources, it is known that traversing "a wide variety of terrains" is accomplished by being able to transition gaits which are meant for each such terrain. Indeed, this paper identifies in the next line of the abstract that it has advanced toward such terrain traversal by using:

a bio-inspired robotic controller able to generate locomotion and to easily switch between different type[s] of gaits.

This furthers the notion that bio-inspiration is relevant here and also bolsters the case that gait transitions are not a solved problem in robotics.

Authors

The Principal Investigator of this paper is Carla Pinto, an associate Professor at the Polytechnic University of Oporto's School of Engineering in Portugal. While this school is not necessarily regarded as a hotbed for robotics research, Dr. Pinto's background is in dynamical systems, a field which, based upon my literature search, appears to have applications in robotics. As principal investigator, Dr. Pinto appears to have had the analytical toolkit necessary to do valuable research, even if her career and her home university are not among the most prolific in robotics.

The other two authors of this paper appear to have been students at the time of its publication. Their credentials are neither a benefit nor a detriment to the credibility of this paper.

Venue

This paper was published at IROS, the International Robotics Symposium. This is a peer-reviewed conference sponsored by IEEE. Given that IEEE is a leading, international engineering organization, this conference seems credible. Additional sponsors that are not associated with IEEE from the most recent conference are: Robotics Society of Japan, Society of Instrument and Control Engineers, Robotics Society of Taiwan, National Taiwan University, and Institute of Control Robotics and Systems in Korea. Because so many international organizations have put their proverbial stamp of approval on the IROS conference, I feel that it is a credible venue.

Further Reasoning

As stated above, Dr. Pinto is not generally a leading researcher in robotics and her University is not among the top names in robotics research. However, this particular paper was published at a conference which was sponsored by some of the bigger organizations leading robotics today. I feel inclined to regard this paper as credible; Dr. Pinto's reputation is not necessarily bad and is more uncertain than anything else, though I feel that we may rely on the reputation of the IROS conference, which appears strong. I also like this paper because it presents a much different approach than the previous papers I have outlined and represents a diversity in the approaches to the problem of gait transitions that exists in robotics right now.

C.1.6) Biology Literature

My First Source

The first biological source I have chosen is [4]. In the abstract, the authors discuss gait transition as a means of managing changing energy constraints and ground reaction forces:

Two studies have focused on potential triggers for the trot-gallop transition in the horse. One study concluded that the transition was triggered by metabolic economy. The second study found that it was not metabolic factors but, rather, peak musculoskeletal forces that determine gait transition speeds.

This is very relevant to roboticists as having a robot change its terrain or somehow otherwise affecting its operating condition(s) imposes a new set of constraints upon the robots motion. Thus having a robot change its behavior as animals do to better meet the demands of these new constraints improves the robot's performance. From this, I feel that this is an appropriate biological inspiration for the question at hand.

Authors

There are four authors listed on this paper, all of whom hold or have held positions at Caltech. The first author of the paper, Steven J. Wickler, passed away in 2007. During his research career, he was the director of the Equine Research Facility at Caltech and the director of Laboratory Animal Facilities at Caltech; the fact that Dr. Wickler held two leadership research positions at a well-respected university such as Caltech adds to his credibility. Google Scholar lists his h-index as 17, which is a respectable figure.

Venue

Both this and the next paper are from the Journal of Experimental Biology. According to the ISI Journal Citation Reports, the Journal of Experimental Biology is ranked 19th out of 76 in terms of impact factor, with an impact factor of 2.722. This puts the Journal of Experimental Biology just inside the upper one fourth of biology journals, which is a respectable rank. However, in terms of total citations, the Journal of Experimental Biology has 20334, which puts it ranked 4th out of 76 by ISI. This is a very high rank and indicates that the Journal of Experimental Biology has great influence on other researchers.

Further Discussion of Credibility

The whereabouts of the remaining three authors of this paper have eluded me thus far. While I haven't been able to track them down, they were all researchers at Caltech at the time, a university whose reputation lends credibility to this paper. While I would certainly prefer to know more about this paper, I feel that the present information is sufficient to regard this paper as credible; the venue, academic associations of the authors, and verifiable appointments of the first author as a leader in research all suggest that this is a valuable contribution to biological literature. Further, I think that because this paper discusses biology in terms of measure ground reaction forces and energy management, it is a very fitting source of biological inspiration for robotics.

Note on Age

While this assignment preferred sources from the last 5 years, I have chosen to include this paper, which is roughly 8 years old. I have done so because this paper explores the energetics of gait transitions in horses, meaning that it explores biology from a physics-oriented perspective that is easily digested by the technical community. In addition, there have been important advances in legged robots since 2003, such as the further development of the RiSE and RHex platforms included in my seed paper; while implementing a gait-transitioning algorithm in a robot may not have been feasible in 2003, it seems more feasible now. Thus, while this paper is nearly 8 years old, I have chosen to use it for this assignment as it presents an area for research that was not necessarily approachable by roboticists until recently.

My Second Source

My second biological source is [5]. In this abstract, the authors note that certain types of gaits appear to form a continuum in horses:

the running gaits (tolt, trot, pace, left and right canters and gallops) could be considered a kinematic continuum

This is highly relevant to robotics as the manner of transition between gaits in animals can be mirrored in robots to potentially endow robots with the same benefits that animals derive from transitioning gaits in this manner.

Authors

The first two authors on this paper, Justine J. Robilliard and Thilo Pfau, have been difficult to track down, leading me to believe that they may presently be students or else no longer involved in academic research. The principal investigator on this paper is Alan M. Wilson, who is currently the Professor of Locomotor Biomechanics and is leader of the Locomotion Research Group at the Royal Veterinary College at the University of London. Given Dr. Wilson's expertise in dealing with biomechanics and his current academic appointments, he appears to be very well equipped to discuss animal locomotion; his credentials lend great credibility to this paper. Further, this paper works directly with animals to examine the features of different gaits and to classify them. This is directly related to robotics as it a biological inspiration for different gaits and elucidates the differences between gaits, thus enabling roboticists to understand when and how gait transitions are possible.

Venue

As with the preceding paper, this paper is from the Journal of Experimental Biology.This Journal's reputation only adds to the credibility of this paper, and I see no problem in using this paper as a source of biological inspiration for gait transitions in robots.

Bibliography
1. G. Clark Haynes and Alfred A. Rizzi, "Gaits and Gait Transitions for Legged Robots", Proceedings of the 2006 IEEE International Conference on Robotics and Automation
2. G. C. Haynes, F. R. Cohen, and D. E. Koditschek, "Gait Transitions for Quasi-static Hexapedal Locomotion on Level Ground", International Symposium of Robotics Research, August 2009
3. Matos, V., Santos, C.P., Pinto, C.M.A., "A brainstem-like modulation approach for gait transition in a quadruped robot", 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2009)
4. Wickler SJ, Hoyt DF, Cogger EA, Myers G., "The energetics of the trot-gallop transition", Journal of Experimental Biology, May 2003
5. Robilliard JJ, Pfau T, Wilson AM., "Gait characterisation and classification in horses", Journal of Experimental Biology, January 2007