How robots can guide us safely through the pandemic

The World Health Organization has recommended that to cut the chain of Covid-19 transmission, people should keep social distance and avoid touching surfaces in public places. The question is: how we can manage our lives in a way that reduces interactions with people? Our answer, as researchers at the Institute of Science and Technology, University of Tartu, is that humans can hand over some tasks to robots.

Fatemeh Rastgar
Me flying quadrotors in the lab. Photo from a private collection

Robots could deliver medicine and transfer patients’ blood tests

In some situations, keeping physical distance and not touching objects is inevitable. For example, in hospitals, doctors and nurses have to check the physical status of the infected, hospitalized people and provide them their medicine and food.

Robots could be helpful assistants for doctors and nurses. In hospitals, without being exhausted, they could acquire information by moving autonomously in patients’ rooms or deliver medicine and food. Moreover, they could help with recording patients’ physical condition and transferring their blood tests.

Robots consist of actuators, sensors, and a ‘brain’

Despite the usefulness of robots, they are difficult to design and program. Generally, there are a handful of important things for researchers to know when dealing with robots, including how robots get information, which data they require for analysis, how they are able to process the received data, and, more importantly, how they can find a path to reach their destination without colliding with walls, objects, and even other robots.

To understand the procedure of the robots’ operation, we should know what parts robots are made of. Each robot consists of different elements, including actuators, sensors, and a brain. Actuators make the robots move, sensors measure and monitor different properties of the surrounding environment, and a central processing unit, which is called the brain, controls all operations and movements. By considering actuator and sensor information, roboticists are able to design algorithms for the robots, making them perform specific tasks. 

Motion-planning algorithms find a collision-free path

One of the challenging tasks for robots is motion planning. Motion planning or path planning is to immediately obtain a continuous path from the robot’s starting point to its final destination, while avoiding obstacles such as walls, people, and objects – for example, how a robot can find an accurate path to deliver food and medicine to patients while it recognizes people, walls, objects, etc. and does not collide with them. Here at the University of Tartu, we develop state-of-the-art motion planning algorithms for different robots.

Generally, the problem of finding a collision-free path that moves the robot from its starting point to the destination is too difficult and, even in some cases, impossible to solve. The complexity of the motion planning problems depends on a couple of factors, such as the robot’s type, size, shape, and position of the obstacles, whether the obstacles move around or are fixed, and finally the environment in which the robot operates.

Robots' motion planning scheme
This is how the robots’ motion is planned. Image credit: Fatemeh Rastgar

Finding a collision-free path requires a lot of code-writing

We try to design algorithms for the robots in intelligent and autonomous manners. Theoretically, to do this, we need to follow these steps. Firstly, we should gather information about the various sensing workspace and record required data, then plan suitable motion-planning algorithms by converting a difficult math problem into an easier one, utilizing different complex mathematical structures, and, finally, find an optimal solution for the problem.

Practically, after solving motion-planning algorithms, the algorithm should be transferred to the robot. The robot’s control center understands the logic behind the problem, processes data, and takes a path without any collisions. At first, we write motion-planning codes and simulate them in computers. Various conditions are simulated, the code is modified so that it runs as fast as possible, and, finally, the results are verified.

The next step is implementing the algorithm in the robots and tackling implementation problems. Also, for this step, verification in different conditions is required. Verification is done by running the algorithms on the computer and implementing  them in real robots. Then, the final result will be that the designed robots are able to move from the pharmacy to the patient’s room in hospitals in a short period of time, while avoiding collisions with people, walls, and other robots. 

Fatemeh Rastgar is a Junior Research Fellow of Robotics Engineering and a PhD student of robotics at the University of Tartu.

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 857518.

This entry was posted in Natural and exact sciences, Research, Studies and tagged , , , . Bookmark the permalink.