SCRS Insights

Title: Biomedical Robotics: The Next Big Leap in Healthcare

“Dr. Shahid Hussain, University of Canberra, Australia, in discussion with Dr. Jagdish Chand Bansal, South Asian University, India.”

Dr. Jagdish Chand Bansal: I am Dr. Jagdish Chand Bansal from South Asian University.
Dr. Shahid Hussain: Yes.
Dr. Jagdish Chand Bansal: This discussion will be uploaded to the Soft Computing Research Society's (SCRS) tech magazine SCRS Insights and will focus on you, your research, and future perspectives. Let’s start—tell me about yourself.

Dr. Shahid Hussain: Yes.

My name is Shahid Hussain, and I am an associate professor of biomedical robotics at the University of Canberra, located in the Australian Capital Territory. I started as an assistant professor and have been working here since 2018.

Before that, I was a lecturer in mechatronics at the University of Wollongong, Australia, for nearly two and a half years. Prior to that, I was an assistant professor at Nazarbayev University in Kazakhstan. I completed my PhD in medical robotics, specifically rehabilitation robotics, from the University of Auckland, New Zealand, in 2013. I also earned my master’s degree in robotics from the University of Auckland in 2009 and my Bachelor of Engineering (Honors) in mechanical engineering with a specialization in mechatronics from the University of Engineering and Technology, Lahore, Pakistan, in 2007.

I have been involved in full-time research and academia since July–August 2013, making it roughly 12 years now. My work focuses on biomedical and rehabilitation robotics, and I lead a research group in this domain, which includes five PhD students and a research fellow. I will discuss collaborations and multidisciplinary research later.

That’s basically my journey.

Dr. Jagdish Chand Bansal: What a diverse experience you have—Kazakhstan, Lahore, Australia, mechanical engineering, mechatronics, and then robotics.
Dr. Shahid Hussain: Yes.
Dr. Shahid Hussain: Yes. My experiences have been mostly positive. As an undergraduate, it was highly competitive—similar to admission into IITs in India. The ranking and selection process were highly competitive because the university admitted only students with the highest grades or GPA in the country. It was a great experience. During my fourth year of undergraduate engineering, I worked on biomedical robotic systems, which led to a PhD scholarship offer from the University of Auckland, New Zealand.

New Zealand had a highly international student environment, especially in the Faculty of Engineering, where the majority were international students. I gained valuable experience working with them, learning about their cultures, and sharing diverse perspectives. It was a fascinating experience as an international student.

When I started my first academic job in Kazakhstan, I didn’t feel much of a difference because the university was newly established, and I was involved in significant lab development. My work is heavily lab-based—I build robots from scratch rather than purchasing off-the-shelf models.

Dr. Jagdish Chand Bansal:  Thanks.

Dr. Shahid Hussain: So that was a big task at a new university. So I was also a part of writing their new curriculum for undergraduate degrees. I also wrote the curriculum for their PhD degrees like the coursework for PhD which the students do. Setting up labs at a new university was a major task. I was also responsible for designing the undergraduate and PhD curricula, including coursework for PhD students. The university followed a Western-style education model, which was familiar to me since I studied in Pakistan and New Zealand—both of which follow a British-style education system

I wrote that curriculum. I established the labs there as well. So that university was built on the model of western style education and since I come from a part of the world. which was once a British colony. So our system was quite similar to what I have been through and I went to New Zealand.

Dr. Jagdish Chand Bansal: Yes, we have common friend also in Najar University

Dr. Shahid Hussain :- Prashant followed me , A friend, who was previously a full professor at Rajasthan Technical University and my senior during my PhD, later joined me in Kazakhstan. I had gained valuable experience there and convinced him to come as well. He arrived a year after me, and after I left for Australia, he continued there.

Since my education was primarily in New Zealand and Pakistan under a UK-style university system, transitioning to Australia was smooth. The Australian university system closely resembles that of the UK and New Zealand, so I faced no difficulties adjusting to job requirements.

At that point, I began developing a new undergraduate biomedical engineering program as a lecturer at the University of Wollongong. This four-year Bachelor of Engineering (Honors) program focused on medical devices—both implantable (such as stents and heart valves) and non-implantable (such as robotic prosthetics, robotic orthoses for rehabilitation, and surgical robots). The program was designed around the theme of medical device development.

Dr. Jagdish Chand Bansal: Actually, this is another area of interest I read about in your publications—biomedical robotics. Can you tell me what biomedical robotics is and how it differs from ordinary robotics?

Dr. Shahid Hussain: Ordinary robotics started in the 1940s and ’50s during the second and third industrial revolutions. In the 1970s and ’80s, during the third industrial revolution, traditional robots were primarily used for tasks like pick-and-place operations in automobile assembly plants. Later, robotic manipulators began performing tasks such as welding and spray painting in the automobile industry.

Dr. Jagdish Chand Bansal: That’s okay.

Dr. Shahid Hussain: These industrial robots did not consider human interaction. If a human operator tried to make changes, the robots were rigid and unresponsive. Their functions were fixed and could not be altered.

Dr. Jagdish Chand Bansal: You mean they were more mechanical and less intelligent?

Dr. Shahid Hussain: Yes, they were very mechanical and rigid. That was the nature of industrial robotic systems.

When we switch to biomedical robotics, these systems must interact with the human body, which presents a greater challenge. For example, if a person loses a limb due to a landmine or accident and requires a robotic prosthesis, the challenge lies in customization. Every person’s walking pattern is unique. With 7–8 billion people in the world, each has a different gait. We must adapt rigid industrial robotic systems, originally designed for tasks like welding and painting, to accommodate the individual needs of each patient.

Another example is designing robots for physiotherapy, particularly for stroke patients. Each stroke is different, and the nature of the resulting disability varies. Even for the same patient, after six to seven weeks of recovery, limb movement and capabilities improve or change. A single robotic system must be able to adapt to the individual disability level of each patient.

Because if we impose industrial robots with the same pattern and forces on the patient's limb during physical therapy, it won't help in recovery. That is the key difference. The shift in biomedical robotics from industrial robotics is toward accommodating injuries and personalizing robots to meet individual needs.

Dr. Shahid Hussain: Industrial robots lack adaptability and personalization for each human subject, whereas biomedical robotics focuses on safety.

Dr. Jagdish Chand Bansal: Yes.

Dr. Shahid Hussain: These robots interact directly with the human body, so they must be extremely safe. Safety is a crucial design consideration in biomedical robotics, along with how the robot interacts with the human body. This falls under human-robot interaction, which is very different from traditional industrial robotics. The challenge is designing robots that adapt their assistance and applied forces to individual needs.

Dr. Shahid Hussain: This is a significant and complex task. Most current research focuses on safety and interaction with patients and human subjects. That is the major distinction between biomedical and industrial robots. For industrial robots, size and weight are not critical concerns, but for biomedical robots, they are essential because they must attach to the human limb.

Dr. Jagdish Chand Bansal: Okay. Yes.

Dr. Shahid Hussain: So, considering the size and weight of biomedical robots is crucial.

Dr. Jagdish Chand Bansal: I was making an analogy with a news report I read in Spectrum a few years ago about small swarm robots injected into the human body to perform cancer surgery from the inside.

Dr. Shahid Hussain: Yes.

Dr. Jagdish Chand Bansal: Is that technology practical?

Dr. Shahid Hussain: Researchers in the US and Australia are actively working on it. Injectable robots for surgical purposes are being developed, though they are not yet commercially available.

Dr. Shahid Hussain: A group at Wollongong University, where I previously worked, was led by a senior professor who designed a robotic capsule.

Dr. Jagdish Chand Bansal: Okay.

Dr. Shahid Hussain: Diabetic patients and others who need regular medication can benefit from targeted drug delivery.

Dr. Shahid Hussain: The idea was to achieve targeted drug release based on location and time. He was developing a robotic capsule that, once ingested, would stay in the body for a few days, releasing the drug as needed.

Dr. Jagdish Chand Bansal: And it adjusts insulin according to the body's requirements.

Dr. Shahid Hussain: Yes, the targeted drug delivery can also be remotely controlled via a phone application. After the drug is fully released, the capsule either dissolves or exits the body. He obtained a Hong Kong or Singapore patent for this innovation.

Dr. Shahid Hussain: He conducted extensive research, but the challenge lies with drug administration bodies. Organizations like the FDA in the USA and the Therapeutic Goods Administration have strict safety and clinical trial protocols.

Dr. Jagdish Chand Bansal: Yes, that’s a challenge.

Dr. Shahid Hussain: Injectable robotic micro-robots must undergo rigorous testing and evaluation before commercialization. However, robotic prosthetics are more commonly used, especially for patients who have lost limbs due to accidents or landmines.

Dr. Shahid Hussain: Many U.S. companies have developed robotic prosthetics for soldiers injured in Vietnam, Afghanistan, and Iraq. These include lower and upper limb prosthetics. Additionally, commercially available rehabilitation robots assist stroke and spinal cord injury patients in physiotherapy, as they are external and easier to regulate in terms of safety.

Dr. Jagdish Chand Bansal: Yes, I was saying…

Dr. Shahid Hussain: Managing safety for external robots is simpler, but ingestible or injectable robots require extensive research and testing before commercial use.

Dr. Jagdish Chand Bansal: This is an exciting technology with tremendous growth potential.

Dr. Shahid Hussain: Absolutely.

Dr. Jagdish Chand Bansal: Biomedical rehabilitation robotics is significantly improving human lives. In my opinion, advancing technology to benefit humanity should be the primary goal of any scientist.

Dr. Shahid Hussain: I completely agree.

Dr. Jagdish Chand Bansal: That's great that you are helping humankind in this. Where?

Dr. Shahid Hussain: In the Western world and also in developing countries, this is becoming a significant problem. We have a rapidly aging population, and you may have noticed in your own community that people often struggle to get up from a wheelchair or bed, or even walk, once they are above 65.

We are developing robotic exoskeletons for elderly assistance. Modern research shows that staying physically active after 60 can help prevent issues like hypertension, diabetes, and even cancer.

That’s why we are designing robotic exoskeletons to help people get up from a chair or bed. These lightweight, small-scale exoskeletons allow users to walk around comfortably. The robot adapts to the user's needs—if they feel tired, it increases assistance; if they are walking fine, it reduces support. Aging-related mobility issues are an even bigger challenge than stroke and spinal cord injuries.

Dr. Jagdish Chand Bansal: I am witnessing the exponential growth of artificial intelligence. Every day, we see new developments in AI. Recent advancements in AI are also benefiting biomedical robotics.

Dr. Shahid Hussain: Yes.

Dr. Jagdish Chand Bansal: As you mentioned, customization is the key difference between biomedical and ordinary robotics. AI-based technology enables better customization.

Dr. Shahid Hussain: Since you brought this up, I have a recently published paper by one of my PhD students on my desk.

Dr. Jagdish Chand Bansal: Great!

Dr. Shahid Hussain: The paper is about a transformer-based approach for predicting transjective energy in neurorehabilitation. It was published in IEEE Transactions on Neural Systems and Rehabilitation Engineering. I am an associate editor for this journal, and your friend Pashan Jamal contributed to this work as well.

So, what we did in this paper was AI-based. Since AI is your area of expertise, starting from your PhD days, a transformer-based approach may not be unfamiliar to you.

Dr. Jagdish Chand Bansal: Excuse me. Yas

Dr. Shahid Hussain: What we are doing here, as I mentioned, is customization based on each patient's needs. You may have heard of regenerative braking in electric cars, where energy is fed back into the battery when brakes are applied. That is the basic concept of regenerative braking.

Now, we are trying to apply a similar principle. For example, a patient who has been out of the hospital for four weeks after a stroke will have significantly more disability and require more assistance than a patient who has been out for three months. The difference between four weeks and three months is substantial.

We initially tried traditional analytical methods to address this but found them ineffective. Now, we are exploring how energy transfers between a robot exoskeleton and a human subject during rehabilitation therapy on a treadmill. Over time, as a patient undergoes physiotherapy, they begin regaining musculoskeletal strength.

Dr. Jagdish Chand Bansal: All right.

Dr. Shahid Hussain: If a robot applies the same force throughout, the patient may not respond well. There is an energy transaction between the human subject and the robot. We spent years attempting to model this using traditional methods, but they did not work. That’s why we applied a transformer-based approach to predict transactive energy.

Dr. Jagdish Chand Bansal: Yes, it is.

Dr. Shahid Hussain: These quantities are difficult to model, so we are using AI-based techniques for prediction. Our next step is to control the transactive energy between the human subject and the robot. AI plays a crucial role here—without it, we could not achieve this level of prediction.

Dr. Jagdish Chand Bansal: Does it require customization over time, or is it entirely based on the patient's movements and activities?

Dr. Shahid Hussain: It depends on the patient’s movement patterns and activities. Now, we are taking it a step further, and our latest research paper is under minor revisions.

Another PhD student applied quantum computing methods to predict neuromechanical control. We are studying a group of patients walking in a robotic exoskeleton on a treadmill to understand how neuromechanics from the brain to spinal cord injury affect robot-assisted gait and upper limb rehabilitation.

Dr. Jagdish Chand Bansal: Okay, wow.

Dr. Jagdish Chand Bansal: That’s good.

Dr. Shahid Hussain: We are analyzing how these factors impact rehabilitation outcomes.

Dr. Shahid Hussain: A few months ago, we implemented another method based on quantum computing. I believe it is more advanced than the transformer-based approach, but we are still in the early stages of applying quantum computing methods.

It was not possible for us to analytically model or predict neuromechanical control of human motion or transactive energy. We were unsuccessful in using analytical methods. As you know, with your mathematics background, Prashant and I are mechanical engineers at heart, so we always start with Newton's Second Law of Motion. However, analytical methods often fail in these cases because human subjects are highly unpredictable in human-robot interactions.

Dr. Shahid Hussain: Modeling human subjects is a significant challenge for us.

Dr. Jagdish Chand Bansal: That’s the main reason artificial intelligence is becoming so popular. Real-world situations, whether human movement or other natural phenomena, are nearly impossible to model mathematically.

Dr. Shahid Hussain: Yes, exactly.

Dr. Jagdish Chand Bansal: The biggest tool for mechanical engineers is Newton’s Laws of Motion, but they only hold true for robotic systems without human interaction. If a robot operates independently, it is not a problem. However, predicting human-robot interactions analytically remains difficult. Maybe in the future, with more advanced sensors and instrumentation, we will be able to make such predictions using analytical methods and Newton’s laws. But at this stage, our sensing technology is still limited.

Dr. Jagdish Chand Bansal: We are running out of time…

Dr. Shahid Hussain: That is also a limitation for us. Sensing technology is not very advanced yet.

Dr. Jagdish Chand Bansal: But this is becoming very interesting.

Dr. Shahid Hussain: Yes, absolutely.

Dr. Jagdish Chand Bansal: So, you mentioned that sensors are another challenge in the development of biomedical robotics. What kind of sensors are you using?

Dr. Shahid Hussain: For position measurement, if we have a robot for upper limb rehabilitation, it mimics the movement of a normal human elbow. We use a position sensor and a joint encoder, which are relatively inexpensive yet quite accurate. We also use load cells and strain gauges. A load cell is essentially a combination of multiple strain gauges.

Dr. Shahid Hussain: We use sensors to measure the interaction force between the human subject and the anatomical joint. We also use electromyographic (EMG) sensors, but they have issues with cross-talk and noise. For example, if we have a motor or actuator in our robot exoskeleton, the electromagnetic actuator generates a current that interferes with the EMG signal. This interference causes cross-talk, making EMG sensors less reliable.

From a sensing perspective, the only reliable options at this stage are position and force sensors.

Dr. Shahid Hussain: However, most of these sensors are only accurate at the joint level. We can measure joint movement, but assessing individual muscle activity in the human body is much more challenging. The data we obtain is very raw.

Dr. Jagdish Chand Bansal: So, apart from AI-based research and biomedical robotics, there is significant scope for research in sensor development as well.

Dr. Shahid Hussain: Yes, for human data measurement, we need more advanced sensors. The challenge is that AI research progresses rapidly, while mechanical engineering research is much slower. Developing new hardware takes time, money, and expertise.

For example, turbo diesel engines gained popularity in the 1980s, and even in 2025, cars still use turbo diesel engines. Over the past 40–45 years, improvements have primarily focused on efficiency rather than revolutionary advancements.

Dr. Jagdish Chand Bansal: The key takeaway from your talk is that the world should not focus only on AI and computer science.

Dr. Shahid Hussain: Yes, exactly.

Dr. Jagdish Chand Bansal: Mechanical and other fundamental engineering fields are equally important. If the new generation does not engage in these areas, AI itself will suffer in the long run.

Dr. Shahid Hussain: Absolutely.

Dr. Jagdish Chand Bansal: At a later stage

Dr. Shahid Hussain: Yes. Take smart and electric cars, for example. AI plays a significant role, but without improvements in motor performance, aerodynamics, and mechanical efficiency, AI alone cannot drive progress. At the end of the day, we need to move things in the real world.

Dr. Jagdish Chand Bansal: You guys…

Dr. Shahid Hussain: If we don't move things, AI will not be useful. It may have applications in banking and data science, but if we want to move something in the physical world, we must improve the mechanical features of that product—such as actuators, sensors, and aerodynamic properties.

Currently, actuators are quite heavy because we are still using industrial robotic actuators invented 40 years ago. We aim to make biomedical robots lightweight, but it's challenging due to the weight of these actuators. Heavy actuators introduce problems related to inertia and other mechanical factors. This is a major challenge for us. Engineering research is needed to improve these aspects, and AI will serve as a tool to enhance these developments.

Dr. Jagdish Chand Bansal: Thank you.

Dr. Shahid Hussain: AI alone cannot solve fundamental engineering challenges. We need better designs for robots, improved sensors, actuators, and overall development in biomedical robotics.

Dr. Jagdish Chand Bansal: Right. Wonderful.

Dr. Shahid Hussain: Yeah.

Dr. Jagdish Chand Bansal: We have crossed the time limit…

Dr. Shahid Hussain: No problem.

Dr. Jagdish Chand Bansal: As an editor of one of the most reputed journals in the world, what is your suggestion for young researchers considering a career in biomedical robotics? Is it worth exploring?

Dr. Shahid Hussain: Yes, it's definitely worth exploring. Biomedical robotics is a multi-billion-dollar market, especially in rehabilitation and prosthetic robots. Based on my academic experience, I would say this is a multidisciplinary field. Young researchers must be open-minded and…

Dr. Jagdish Chand Bansal: Heat.

Dr. Shahid Hussain: I'm open to new suggestions. For example, I usually meet physiotherapists every second day for my research, and I'm also developing a long bone femur fracture reduction robot.

I regularly meet orthopedic surgeons as well. With my background in fundamental mechanical engineering, I initially took a conservative approach, thinking that I could follow a specific methodology. However, the problem is that when you talk to end users, like physiotherapists, patients, or orthopedic surgeons, they have a different perspective on the problem you're working on.

Anyone pursuing a career in biomedical robotics should know that the starting point is not just your engineering background. You can't build things based only on what you know. You need to build things based on what the end user requires. It’s funny, I used to hear that the relationship between an architect and a civil engineer is complicated, but now I realize it’s true for us as biomedical robotics researchers too. Our relationship with healthcare professionals, such as surgeons and physiotherapists, is complicated because their expectations for the functionality of our robots are different from what we think they should be.

At the end of the day, you have to work based on their input and feedback. Then you can explain, “This isn’t possible from an engineering standpoint,” if necessary. Often, they listen well and are open to your thoughts. My advice is to put your engineering knowledge aside and approach the problem with an open mind, asking end users what they truly need. Their needs are often very different from our initial thoughts.

Dr. Jagdish Chand Bansal: Exactly. So, you mean to say that it's not necessary to have a background in a particular engineering branch to pursue a career in biomedical robotics? It’s all about exploring the subject?

Dr. Shahid Hussain: Yes, exactly. In my research group, I’ve supervised PhD students from different backgrounds. Some were mechanical engineers, some had degrees in robotics and mechatronics, and one was a materials engineer working on lightweight materials for robot exoskeletons. I also have two applied mathematicians in my group.

Dr. Jagdish Chand Bansal: That's great to hear.

Dr. Shahid Hussain: Yes, this paper written by Navidid Khan, an applied mathematician with a master's in applied mathematics, is one of the contributions from my research group.

Dr. Shahid Hussain: is not an engineer, but they do a wonderful job as well. This is not limited to anyone.
Dr. Jagdish Chand Bansal: So good.
Dr. Shahid Hussain: It depends on the nature of the research you want to do. I have worked in this field for 16 years since I started my PhD. I know which project can be done by which person. As I said, it's a multidisciplinary field. We are now getting people with physiotherapy degrees as well to do research in this area with their PhD.
Dr. Shahid Hussain: Some people are coming in June from overseas with physiotherapy degrees to do a PhD in rehabilitation robotics or biomedical robotics. Your very specific undergraduate degree is also not required.
Dr. Jagdish Chand Bansal: So, it means a particular background doesn't significantly affect the study in biomedical robotics.
Dr. Jagdish Chand Bansal: Thank you, Professor Shahid Hussain.
Dr. Shahid Hussain: Thank you very much.
Dr. Jagdish Chand Bansal: It was an exciting discussion, and I’m sure the readers of the transcript will learn a lot. I’m also sure the world will look for technological development not only related to artificial intelligence and computer science, but also from the electrical, mechanical, mechatronics, and robotics points of view. The overall development is required for the benefit of human society.
Dr. Shahid Hussain: Yes.
Dr. Jagdish Chand Bansal: Thank you, Professor.
Dr. Shahid Hussain: Thank you for your time. It was really nice talking to you.
Dr. Jagdish Chand Bansal: Thank you.
Dr. Shahid Hussain: I will stay in touch with you later. Thank you. Bye.
Dr. Jagdish Chand Bansal: Thank you. Bye.