How GlobalLogic helped Haystack Robotics build a UV disinfection robot

In hospitals, hygiene is essential. But the cleaning process, which is time-consuming and difficult to scale, remains manual labor. This often means that the process is also inconsistent. Healthcare facilities are under pressure to care for more patients with fewer resources. These challenges are why GlobalLogic’s Silicon & Embedded division helped Haystack Robotics design a mobile robot to handle disinfection.

The project began during the COVID-19 pandemic, when disinfection became a critical use case for robots.

“Obviously, this was a great use case at the time, in terms of being able to disinfect hospitals and healthcare facilities,” said Kunaal Mahanti, head of physics AI at GlobalLogic The robot report. “AS [Haystack] he had the idea of ​​being able to perform disinfection better.”

“Even though the disinfections were done manually, the coverage was not really good,” he added. “We were really the engineering and collaboration partner there to see how we can take this concept and make it happen.”

The result is an autonomous mobile robot (AMR) that combines real-time perception, edge-based intelligence and navigation to provide hospital-grade UV disinfection with minimal human intervention. The Santa Clara, California-based company said it can operate safely alongside people, interpret its surroundings and adapt to changes in its environment.

GlobalLogic aims to create a scalable platform

When designing his robot, Mahanti said GlobalLogic didn’t want to simply quickly design a system for one use case or environment.

“We wanted to make it an expandable platform,” he explained. “The idea was that you could look at, for example, the part of the track, which is the thing that the robot is carrying, independently of what it’s carrying, which is the UV. Architecturally, you keep them separate so you can mount other things later.”

“It went from a concept to what would be an architectural foundation, so it’s not just a product,” Mahanti said. “It effectively becomes a class of solutions that we can build for the same industry segment.”

While GlobalLogic hasn’t done many other hospital deployments, Mahanti said the company has used some of the UV robot’s core components for other projects.

“When you look at the platform moving, there are things that have emerged from there in terms of motor control,” he said. “So we use ROS to do the programming and the sensor fusion part, which is reusable, and we’ve used that. So if not necessarily the same device, but that architecture in general.”

Addressing the challenges that arise from hospitals

Healthcare facilities are crowded, chaotic and involve many closed spaces, making it difficult for robots to work.

“The known part was that, obviously, you have to interact with humans. The unknown part is that there are doors, there are things that need to be opened and you can’t always break in,” Mahanti said.

This is different from the most common use cases for mobile robots in the warehouse or manufacturing, which typically involve easy-to-navigate open spaces

“There are two modes that we ended up creating for [the robots]” Mahanti noted. “Essentially, one was this autonomous mode, where it navigated and found the right room. The other was the “follow me” mode, in which case he would follow a caregiver or nurse.”

This means the assistant can open doors and ensure spaces are ready for the robot to enter. This also means that a patient would not be taken by surprise by a robot entering their room, for example.

However, designing an AMR that can operate in these complex spaces has created advantages in other areas of GlobalLogic’s business. With a much higher level of security, the company could apply the platform to other industries.

“The other learning is just being able to be in an environment where you can’t raise your voice, for example, when we design engines or look at what we need to put in [the robot]” said Mahanti. “These things are specific to this industry, but addressing them has helped us build higher fidelity machines elsewhere, where it’s not necessarily a given, but it’s a good thing to have. Why wouldn’t you want to have something nice and quiet that doesn’t interfere with the rest of the environment?”

What’s stopping the adoption of robots in hospitals?

While more and more robots are making their way into hospitals, most people still can’t walk into their local hospitals and find robots roaming the hallways. This is still an emerging area for robotics, so what challenges still exist? Mahanti said three things come to mind.

The first is the cost. “There needs to be an ROI,” he responded. “For example, if you look at this use case during COVID, you needed fewer people in hospitals at that time. You had to be safe. It was a big deal. So the ROI was justifiable.”

The second is compliance with medical standards, which becomes much more complicated when using robots that will interact with the public and patients.

The third is the perception of robots. Mahanti acknowledged that hospital staff don’t want a large, bulky robot taking up space in busy hallways and elevators.

Prem Balasubramanian, chief technology officer at GlobalLogic, said secure integration into hospital workflows is also challenging.

“There are clearly safety concerns, from multiple angles. Navigating a very crowded situation is not easy,” he said. “For a disinfectant use case, if a robot waits for people to keep moving and then moves slowly, that’s fine. But if you use it for anything else, like if a doctor calls and you need to get something quickly to help a patient or carry something, a robot might get stuck because it sees people moving and can’t move.”

The hospital is difficult in itself, and even more so when you consider how much they are made for human beings, Balasubramanian said. For example, a robot might be able to navigate an emergency room in perfect conditions and without obstacles, but the addition of patients constantly moving or being moved and staff running around to solve critical problems makes navigation much more challenging.

“The entire workflow inside a hospital is built for humans. You can’t just make a robot think it’s a human. So, there’s work to be done to convert some of the workflow, the spaces, everything,” Balasubramanian said.