Scientists Create a Mathematical Model of Insect Vision for Mobile Robots

A team of scientists from South Ural State University has created a mathematical model of a complex insect eye, which will later be used for a prototype device based on computer vision technology. In the future, these developments will make it possible to create smart mobile robots to deliver medicines to people who do not have the opportunity to go to the pharmacy on their own. The study was published in the highly-rated scientific journal Mathematics (Top-10 Web of Science).

The development of devices with computer vision is a very relevant and rapidly developing area of scientific research. Chinese scientists, for example, are already actively using such devices on drones. Junior researcher of the Department of System Programming Artem Starkov, under the guidance of Doctor of Mathematics, Professor Leonid Sokolinsky, created a unique two-dimensional mathematical model of complex vision. A new mathematical model created at SUSU will make it possible to solve complex problems of navigating autonomous mobile robots on the ground and in the air. The insect compound eye made up of ommatidia, tiny independent photoreceptive units, inspires it.

"Insect eyes have a fundamentally different structure than human eyes. There are no moving parts, and a fly, for instance, has to turn around to see a static object clearly because its eyes react to movement. Our model is the first and so far, the only one that describes the planar binocular vision of the compound eye while making it possible to measure the distance and azimuth to surrounding objects. Video sensors similar to the compound eyes of insects are a promising alternative to digital cameras. Such video sensors do not have moving parts and do not require any control," Leonid Sokolinsky explains.

The model provides a necessary and sufficient condition for the visibility of an object by each ommatidium. On this basis an algorithm for generating a training data set is compiled to create two deep neural networks: the first detects the distance, and the second determines the azimuth to the object. The experimental results showed that the proposed method determines these parameters efficiently and accurately. These and other studies will make it possible to create robots that can work autonomously without any human intervention. They can be used on the ground and in the air to navigate and reach their destination, bypassing obstacles.

"We live in a world where technology serves people, ensuring their comfort and safety. Imagine how a robot delivers medicine to a sick person who cannot leave the house. A robot opens a special box installed on the windowsill from the outside, and then a person opens it from the house. The robot leaves a set of medicines that a person needs, making his life a little better, a little longer. This is the task of science," Leonid Sokolinsky says.

Shortly, together with the Department of Optoinformatics, it is planned to create a prototype device using the developed technology.

The study is being implemented within the strategic project Intelligent production. This project is one of the priority areas of the Priority 2030 program. South Ural State University is a university of digital transformations, where innovative research is carried out in most priority areas for science and technology development. Following the strategy of scientific and technological development of the Russian Federation, the university focuses on the development of large scientific interdisciplinary projects in the field of the digital industry, materials science, and ecology. In 2021, SUSU won the competition under the Priority 2030 program. The university performs the functions of the regional project office of the Ural Interregional Research and Educational Center.

Provided by South Ural State University