Autonomous Robots are incredibly complex systems dependent on several environmental factors that require sophisticated approaches to mathematical modeling !
Ameya Salvi, our next pathbreaker, works as Robotics Researcher at Hitachi America Ltd., a global multinational which is at the forefront of technological innovations.
Ameya talks to Shyam Krishnamurthy from The Interview Portal about his PhD on Skid-Steered Robots (such as army tanks) that use high fidelity simulators (also know as digital twins) for creating studies and gathering data (data such as sensor data like camera, vibrations, etc) to design them.
For students, learn something because you really want to, because learning for the sake of expanding knowledge is really joyful in itself.
Ameya, can you share your background with our young readers?
I was born and brought up in Mumbai, and I took a pretty standard academic route of taking the Maharashtra State board examinations for secondary and higher secondary education, and then the Maharashtra state’s engineering entrance examinations. I would say I did quite well in school examinations (I was never a rank holder but was slightly above average in my academics). From a very young age, I was always interested in science, specifically in Physics, that slowly shaped my career decisions. Of course, my knowledge about what exactly I wanted to pursue in science was quite limited and it was only later (after my graduation) did I understand clearly about my interests.
My parents were in the administration departments of Bombay Municipal Corporation and BEST bus services (father and mother respectively). They held degrees in Arts and Commerce respectively, and they always encouraged me to be serious about my education (a pretty common occurrence in all Indian families).
What did you do for graduation / post-graduation?
I did my bachelor’s in engineering (BE) in Mechanical Engineering from University of Mumbai and then completed my Masters in Science (MS) in Mechanical Engineering, followed by my Doctorate in Philosophy (PhD) in Automotive Engineering, both from Clemson University.
What were some of the key influences that led you to such an offbeat, unconventional, and unique career in DeepTech Research?
My interest in Physics made me opt for Mechanical Engineering as a choice for my undergraduate education. While I didn’t clearly understand what within mechanical engineering was of keen interest to me, I did broadly like the study of motion of bodies (what we know as general mechanics) and wanted to study its engineering applications. During my Bachelors, while I was intrigued with the coursework, It was never quite enough for me. I wanted to experience what it is like to design and develop mechanical systems, an interest that led me to co-found/revamp a formula student team in my college (a student level competition where students design and build a racecar to compete at national/ international level : Formula SAE ) – a memory I am going to cherish for a lifetime.
It was during my F-SAE days that I realized that I needed a much deeper understanding of what it takes to build complex mechanical systems, how research needs to be conducted and what different areas of specialisations can I take up. It is this experience that led me to think about post-graduation (MS/MTech) as a career option instead of applying for jobs in the industry.
While I started my MS application process, I worked briefly (for roughly a year) in an electric vehicle startup, as a mechanical engineer.
As a mechanical engineer, I was working for an electric motorcycle company (That was named Zeuva at that time, unsure if it is still around). I was part of a small team and my job was to assist the design, development and integration of a modular battery packaging/ cooling system for the lithium-ion battery packs for motorcycles. While the company already had a preliminary prototype, we were working on refining it with some novel R&D.
During my brief work tenure in Mumbai, I realized I severely lacked critical knowledge required to build a system from ground up. I realized I needed to be part of a larger team and needed guidance and supervision to carry out critical research. This experience further strengthened my ambition to pursue further education to improve my domain expertise. Working over here made me realize that pursing systematic research is quite involved and something that is seldom covered during Bachelors education – a gap I severely needed to fill.
How did you plan the steps to get into the career you wanted? Tell us about your career path.
My journey towards my current position was the result of personal decisions and some necessary steps due to unavoidable circumstances. I joined Clemson University for MS in Fall of 2018 and at that point, my only interest was to pursue MS, take up a job for a few years and with the experience, return to India to contribute to the Indian product development and R&D sector. During my MS, I fortified my understanding of some of my core interests in the study of motion by studying topics such as Dynamical Systems and Controls.
I realized robotics as a core research area would be a suitable option to apply my knowledge of dynamical systems and controls. Through projects and coursework, I strengthened my theoretical understanding of how systems worked and applied them on hardware and software projects for fortifying my knowledge base.
In an attempt to strengthen my fundamentals as a researcher, I worked as a student researcher on an MS thesis topic where we were exploring the idea of using the thermoelectrics technology for powering wearable devices. The essential idea of the research topic was to harvest body heat to power devices such as watches and other wearable devices.
While the subject exposed me to the concept of dynamical systems (a key foundational subject in controls and robotics) it never really allowed me to explore control engineering and robotics, which slowly dampened my motivation to pursue the subject. As a result, I slowly drifted away from the topic and later focused my attention on coursework on control and dynamical systems.
While this experience wasn’t directly related to the robotics work I wanted to pursue, it exposed me to the fundamentals of pursuing research such as conducting literature studies, benchmarking solutions and analyzing data. This knowledge was transferable to any research domain I was going to pursue in the future.
During my MS, I also gained some practical experience in Proterra, an Electric Bus company in their Thermal and Control Systems team. While I didn’t exactly get the ‘robotics’ experience that I wanted to, the internship helped me understand the inner workings of an engineering team, especially when putting together a large system such as an Electric bus. I was able to apply my knowledge of the theory of Dynamical Systems during one of the projects where I was appointed to design a HVAC (Heating, Ventilation and Air Conditioning) system. One significant learning experience at Proterra was, how critical it is to strengthen your fundamental principles in Engineering and Mathematics as they can be applied to an extremely large set of physical challenges.
After my summer internship and MS, I was eager to graduate and join the work force in May 2020. Unfortunately, that summer was gripped by the Covid19 pandemic, which brought forth significant challenges in getting a research-oriented job.
How did you get your first break?
Fortunately, during my job search, I was networking with several of my professors and teachers for potential research positions/ lab assistant positions that would help me start earning (at least to cover my living expenses). In one of such interactions, I was offered a PhD research position in the Automotive Engineering department of Clemson University. While PhD was never on my radar, I realized that this research position perfectly aligned with the kind of experience I wanted to gain. Furthermore, it would be a paid position, covering my tuition with supplemental stipend. I would always be obliged to the professor (who later ended up being a Co-chair on my PhD thesis committee and with whom I worked very closely for the next few years) for this opportunity, which I could consider was my first break in the research field.
The title of my dissertation was “Learning enhanced system identification and control of skid-steered wheel mobile robots”. To give a brief background, my research was funded by the US Army’s Ground Vehicles Systems Center. The US Army was looking for ways to enable autonomous vehicles in their operations and was looking for academic collaborations to conduct research and development for improving those capabilities. To this end, me and several other of my colleagues were part of what is know as the VIPR-GS center at Clemson University. All of us were researching slightly different topics, amongst which, my topic was related to investigation of the motion planning and control of skid-steered robots.
Skid-steered robots (such as army tanks) are vehicles that do not have a steering mechanism and their turning motion is executed by a speed differential between the right and left side wheels (if the right wheels turn more than the left wheels then the vehicle does a left turn and vice versa). As a result, motion of these vehicles is dependent on several environmental factors such as vehicle dynamics, tire slipping, soil characteristics and terrain roughness. Incorporating all these different factors in a mathematical study can be very challenging and thus what I wanted to investigate was if I could use some contemporary tools such as machine learning and reinforcement learning for this particular investigation. More specifically, I was using high fidelity simulators (also know as digital twins) for creating studies and gathering data (data such as sensor data like camera, vibrations, etc) that I would later used for designing my autonomous control framework for the robot (a technology that would allow the vehicle to move on its own, aka driverless/self-driving motion).
To be completely honest, finding a job is a mixture of several factors, some within your control and some that are out of sheer luck/timing. Around end of 2024, I realized that I would be able to finish my research in the following year and so I began my job search. Since this is something I was doing alongside my research, I wasn’t able to dedicate a lot of time to it and hence it took me several months to finally find a job. I used to reach out to several people such as my seniors from the university or personnel from the industries that I was interested in working for. I personally was open to Academia as a career option as well, and hence was open for a suitable post-doc position too. Online applications and networking helped me in not only getting a few interviews (which didn’t convert to job offers) but also allowed me to gain valuable experience regarding how I can present my research to someone from industry or a widely different research field. This experience was very crucial as it prepared me for future interviews, like the one with Hitachi that eventually converted into a full-time position. With Hitachi specifically, I believe it was more about timing and our (mine and company’s) interests aligning at the right time, that probably made me a suitable candidate.
What were some of the challenges you faced? How did you address them?
While there were several challenges, I would like to share my learnings from these challenges, which were the critical takeaways of my PhD experience.
Understanding how academic research differs from industry research.
My experience in the industry has taught me to look at every problem from a utility perspective (how can I make this cost efficient, how can the company derive profit, etc). Academia on the other hand approaches research from a knowledge point of view. The objective of academic research is to generate knowledge that can be later utilized by several others to solve their specific challenges.
Research is a marathon, not a sprint
PhD can be exhausting. We are typically working 24×7, without much vacation, for anywhere between 3-5 years. This can take a significant toll on your life, especially when experiments are stuck and no progress is being made. Often at such times, one has to calm oneself down, and refocus on why he/she began this journey in the first place. More critically, one has to become process oriented and start enjoying the process of conducting research without being desperate for getting results. I believe such an attitude is helpful in several walks of life.
The most critical outcome of your PhD is you, not your research. Often students associate their success with how many papers they published or patents they filed. More critically, the truest product of a PhD are the students themselves that are going to go out in the research world and apply their knowledge and skills in the future. The end of a PhD is thus really the beginning of the Research career and one has to understand that the best product that came out is an independent researcher (which is you!)
Where do you work now? What problems do you solve?
At present, I work for Hitachi America, in their Research and Development Division. My job title is Robotics Researcher and I spend most of the time during my day pursuing research similar to what I did during my PhD. The primary difference is that in my job, my research has more specific goals and is not as open ended as during my PhD.
What skills are needed for job? How did you acquire the skills?
The key skills necessary are:
- Ability to identify what a true technical problem is. Your critical thinking skills are really tested here.
- Knowledge of current state of the art technologies, so that you can come up with novel solutions to challenging problems in minimum time.
- The ability to communicate your ideas. You often work in interdisciplinary teams that may not share the same level of technical expertise in your research subject. You should be able to communicate your ideas to them.
While a lot of these skills are developed during your PhD, it is a constant learning process to upskill yourself.
The thing I love most about this job is that it allows me to stay curious and constantly keep on increasing my knowledge base.
How does your work benefit society?
Hitachi America, a global multinational, is at the forefront of technological innovations. While I personally may not be impacting the society, I like to be a part of the larger system that can create an overall impact.
Tell us an example of a specific memorable work you did that is very close to you!
While there is no one particular work that is very memorable to me, I would say all times I worked in teams was an extremely fun experience for me. I believe, as a group we can do much more than as an individual. Every time I was a part of larger teams, I felt a sense of camaraderie and a much joyful overall experience.
Your advice to students based on your experience?
Learn for the sake of learning. While you may be tempted to gain degrees and certifications as a stepping stone towards a lucrative job, it will only harm you in the long run. Learn something because you really want to, while convincing yourself that learning is really joyful in itself. Such an attitude towards learning makes the process a lot easier, makes you a lot more resilient and helps you retain whatever you have learned for a very long time.
Future Plans?
I am presently happy where I am. Perhaps like almost every NRI, I too feel like coming back home and contributing to the Indian Research and Development initiatives, but I am not too sure about it at this point. Hopefully sometime in the future, when I find a suitable problem to tackle, I might take on that challenge.