Note: This is a post from my old blog. I am reposting this here to get catch a drift of jekyll et cetera.
Recently, I’ve been questioning my plans for the future and after a certain amount of reflection, I came up with the following (seeming worthwhile?) reasons why a grad school education would be worth the shot.
Passion for conceptualization. I fancy conceptualizing, finding issues and directions, definitions, exposition and critical insight. In developing a new model or notion, I love making definitions and then developing algorithms and analyses which ultimately lead to the solution. The opportunity to feed (no, feast!) on ideas – to investigate the physics behind machines and to understand the interconnectedness of components, hearing people’s opinions, seeing the flaws in their ideas and improvising upon them to achieve some sort of epic grandeur has always been an intrinsic component of my thought-actions.
Quest for challenge. I like to try things that other people find difficult. Yes, I might spend my free time differently than most, but I think synthesizing disparate sets of data and making robots fly, track objects and help humans live better is sexy. Also, I’m a pretty thoughtful person and I have a great respect for people who know a lot of stuff about an esoteric subject. It helps that I have long fancied tech gurus and nerds and have many times imagined myself to be some rock-star inventor someday. Every so often, I would listen to a google/TED talk host or a similar facsimile and the sheer majesty of watching creativity in display always awed me. I fancy myself one day, at one of those Google tech talks varnishing the truths of some newly coined robotics technology. It also helps that I like being a disseminator of knowledge and I figure that a PhD is a prerequisite for me if I am going to be someone that would matter in academe in the future.
Evolution from traditional mercenary work Since finishing my BS degree almost a decade ago, I have worked in the private sector for various international companies. I know I can hardly work on a problem that I want for myself without ensuring it is a problem that the business is interested in or one that could bring in immediate results in terms of profits. I also know the pressure that comes with delivering according to set goals by business executives - goals that sometimes make absolutely no sense to me except that they rake in profit for the company. Sometimes, the means to achieving such goals could be immoral and oftentimes unethical. But hey, so long as no one catches you and you are delivering results, who cares? My experience to come back to academe stems from my work experiences as well as love for learning, innovation and other life experiences. I imagine that the life of an academic provides an eternal exploration for creativity and comfort in life. For example, I have been working as a lecturer at a small state college this past one year and I get so much satisfaction from my job. I can work around my schedule without disrupting someone else’s almighty timetable or Gantt chart as would be the case in a company. I also get to explore in my little world the possibilities of technology if one applies the right tools. I know the limitation placed on creativity with 9 – 5 jobs and I am not impressed with this sort of job at this stage in my career. Someday, I don’t plan on having a boss. I also don’t plan on having a 9 - 5 job. Given my love for/interest in higher education, a Ph.D. seems like a good tool to have to achieve this goal. I certainly would not mind becoming a rock-star professor but I am also in this PhD endeavor for the freedom and the opportunities. So rather than serving as a paid-agent in propagating someone else’s vision (as in a company rote-job), I would love to pay for the chance (with bloody sweat and hardwork) to solve open-ended problems. Opportunity to develop my skills to the uttermost. In a PhD, I can choose what skills to develop, for example, coding, teaching, innovating (certainly) or even delivering a talk. I know that in exploring questions, I will have access to experts in their field. I want to develop a sense of confidence in the power of rational thought and the way of approaching methodically the job of solving complexity. This, I believe, would constantly take me to the edge in my quest to utilize my talents to the uttermost as I bring solutions to otherwise unsolvable problems. I expect I would be able to research anything and have the proclivity to question and understand all that is around me and seek out new ways of doing it or seeing a problem. Because I love asking difficult questions and finding the right solutions, I expect that the aftermath of the PhD pursuit would give me the confidence that I can move around research areas, pick shit up quickly and say something enlightening about it, even if others have looked at it a long time. Pursuing a PhD is fun itself: The opportunity to have an increased appreciation for creativity in other people and every area of life, to think creatively and seek it out and to be more appreciative of what it took to do creativity and how it is different from previous works, to enjoy a new sense of taste and a critical sense, willingness to contradict conventional wisdom and being constructively critical, for me outweighs the costs (limited earning potential compared to my colleagues in the industry etc). In the unlikely scenario that I do not use my PhD in my future career, it will have been fun and a world-class training that is priceless not to mention the next five years would be a productive few years of my life. I know that it won’t be a rosy picture all through graduate school: papers might/possibly will get rejected; what I work on may not interest much people outside of my little academic circle as the return-on-research value might not come early in my PhD pursuit or possibly even during my PhD program but I know the frustrations will be a part of the learning process: honing my research thrusts and learning how to achieve a high-level research objective along the way. I hope that it would help that I have learnt a great deal from knowing what it’s like to fail, come second, let people down, to let myself down in the past. Because success is a mix of natural talent, hard work, judgment and, yes, a bit of luck along the way (which I guess I have copious amounts of), I am confident that I will be a good doctoral candidate.
WHY ROBOTICS? My decision to study robotics was a culmination of my experiences in an eclectic field. During my MS degree I happened to take a module titled “Robotics and Multisensor systems” which expanded on topics such as robotic manipulators, robotic architectures, robots geometry and kinematics, applications of robotics and this gave me a fair understanding of basic theoretical principals and research challenges in robotics (designing assistive rehabilitation robotics, medical robotics, trajectory path planning and SLAM among others). But more importantly, the multisensors part of the course introduced me to target tracking and computer vision applications to robotics, data fusion models, control and estimaton and process architectures and after a couple of labs and self-reading of many texts, my interests in the applications of robots boomed. I consulted journals and technical publications to gain a better overview of the state of the art research and current research challenges. When it came time for my MS cohort to choose project topics, I decided for robotics having enjoyed the basic principles of robotics and being willing to accept the challenges of conceptualizing new models to tackle challenges in robotic development. With my advisor, Dr Tony Dodd, who directed the autonomous systems research lab at the time, I came up with the proposal that I develop a machine vision based navigation system for a UAV. This project work taught me how to approach complex engineering tasks. Developing a new flying robot alone required an immense amount of consideration: from having to carefully choose the hardware and mechanical components that could aid achieving a light, flying object. The first time I coupled my arduino-based quadrotor and tested it for flight, I soon found that government regulations in the UK against outdoor flying objects and environmental perturbations would limit the scope of my experiment to an indoors one. I joined a PhD pet project tasked with developing a waypoints controlled quadrotor at the robotics challenge competition in Eindhoven to gain more experience at the subject, read technical papers, talked with other people doing similar research and consulted publications with related research topics. I ended up limiting the scope of the project work to an indoor environment where flight path was predetermined and goal definitions for the project was redrawn to include visual-guidance to a pre-determined landing target, image-based assessment of target and subsequent landing. Doing this project taught me more than how to build a visual-servoing robot. I honed my control theory skills to fit the definition of my research goal, and more importantly, I learnt the value of having a good advisor who could guide me in the early stages of my research development. I also learnt about planning and coordination involved in building a large, 6-DOF system with a team of people of varied skills: departmental technicians, colleagues with robotics expertise, PhD robotics researchers and mechanical engineers. I learnt to have an increased appreciation for creativity in other people and every area of life; My experience helped me develop a new-found sense of appreciation for hardware and application-based research and gave me the confidence to commit myself to more ambitious dreams. Though I enjoyed Robotics, I increasing found myself wanting to solve more relevant problems. Once while attending an engineers without borders symposium at my school, I learnt about a project challenge of developing a low-cost wind turbine for Haley farm in Sheffield and how these talented students were working on assistive robotics for disabled children and sick people in the South Yorkshire region. I soon enrolled with this team and began to combine my scientific elegance with practical impact to a greater degree. I soon found myself raising funds for the assistive robotics program along with the project director. Within two months, we raised a few thousand pounds which was enough for us to put together a perfect team of engineers that successfully developed a semi-automatic prosthetic limbs for a boy who had walking deformities. Also, by speaking with university authorities, we gained scrap metals which we melted and moulded in the materials engineering department to build the engine for our wind turbine. Although generating a measly power of 700watts, we were proud of our little progress and our works brought us a little fame and fortune enabling the “Engineers without Borders” society to be named the University Society for the year. I found the experiences invaluable as I had the chance to work with engineers from diverse fields including mechanical, and civil and electrical engineering and all the more reinforced my cross-disciplinary research interests. I learnt that research areas are in a constant state of flux; as new technologies and problems appear yesterday’s problems are solved and go out of fashion. Old problems resurface in new clothes and lead to wonderful new discoveries.
What, if any, specific topics/problems in Robotics are you hoping to solve/investigate?
My experience with vision algorithms in autonomous systems so far has confirmed the importance of finding efficient distributed vision sensing algorithms for multi-target tracking by a team of cooperative robots, dynamic map-building by cooperative agents in a leader-follower relationship, distributed formation of a group of mobile robots or synchronization and rendezvous as in consensus networks. I therefore want my graduate research focussed on developing better build systems (efficient control and estimation algorithms for energy-saving in nodes through embedded systems processing), error checking mechanisms (by exploring control algorithms for optimum error correction and data compression models in control unit applications) and embedded video processors capable of opportunistic sensing, efficient navigation and scene analysis by autonomous robots. As my research carries along in graduate school, I look forward to developing effective communication protocols for the efficient delivery of sensed information in energy-constrained environments. Essentially, in a high-level sense, my research goal is to use visual-servoing and computer vision techniques, with minimal computational costs, to improve reliability and accuracy in robotic control and target tracking applications