Phytomodel: digital counterparts of city farms will help grow fruits with a given taste

Russian and Belarusian scientists have developed and tested a digital twin for smart farms. This computing system predicts environmental changes with almost 100% accuracy. When implemented into the control loop, it will help maintain optimal ecosystem parameters close to optimal ones. This can be achieved using simple, commercially available Internet of Things devices: smart sockets, sensors, and servers. The development was tested on a vertical city farm with 4,500 seats. Further development of the technology will make it possible to create fruits and vegetables with a programmable elemental composition and a "designer" taste. For more information, see the Izvestia article.

How to improve the environment for plants

Scientists from Siberian Federal University (SibFU) and the Center for LED and Optoelectronic Technologies of the National Academy of Sciences of the Republic of Belarus have developed and tested a digital twin with artificial intelligence for phytotrons. These are chambers where plants are grown in conditions of full climate control, without outside access to heat and light. For example, in urban farms that are located indoors or underground.

The developed mathematical apparatus is a system of equations that is able to predict changes in the environment in the chamber with almost 100% accuracy under certain control actions.

— A digital twin is a dynamic virtual computing model that shows the current state of a smart farm. It is generated using data that is received from internal sensors. Thanks to the work of artificial intelligence and analytics, the digital twin can predict events, simulate scenarios and offer the most profitable solutions for growing a particular crop," Ivan Timofeenko, head of the SibFU City Farming Laboratory, told Izvestia.

Photo: IZVESTIA/Sergey Lantyukhov

He explained that the research was not conducted in a laboratory setting, but on an operating city farm. The company is equipped with seven-tier racks for 4,500 seats and produces hundreds of kilograms of products per month. The equipment includes powerful LED lighting, air conditioning, heating, dehumidification and other elements for environmental management.

According to the scientist, the statistics collected at the facility became the basis for training neural networks. As a result, a model was built that allows for virtual experiments with a phytotron digital twin. This makes it possible to change the operating parameters of the equipment and assess their impact on environmental conditions: temperature, humidity and other factors.

— The results of the work showed that, for example, during the virtual simulation, the temperature deviated by no more than 0.1 °C from the real values, humidity — within 2%, and the forecast of energy consumption coincided with the actual data by 99.5%, — said Ivan Timofeenko.

Photo: IZVESTIA/Eduard Kornienko

The city farm management system is based on simple, commercially available Internet of Things solutions - smart sockets, sensors, and servers. Thanks to this, the digital twin is useful not only in scientific experiments, but also in the daily practice of urban farms, he noted.

"A class of consumers of the domestic component base has matured in the country"

Russian physicist Sergey Ivanov — on advanced domestic developments in the field of optical microelectronics

How to program the properties of vegetables and fruits

Achieving optimal environmental parameters in the future will allow experiments on plant property management. For example, you can get tomatoes with a designer taste or with a complex of useful elements.

— In one of the studies, we encountered a problem: there are types of lettuce that are genetically supposed to be red, but when grown in normal white light on urban farms, they remain green. As a result, we were able to choose a wavelength that, when turned on at a certain moment, creates stress for plants. This does not greatly affect the set of biomass, but it allows you to change the color of the salad," Ivan Timofeenko gave an example.

Photo: Global Look Press/Wilfried Wirth

However, plants are complex living systems, and modern science is still far from creating their full—fledged digital counterparts, he stressed.

— The use of digital twins for the tasks of increasing yields seems to be effective. In particular, the accuracy of the models makes it possible to optimize energy consumption management. Such technologies are especially important in the long term, when stable feedback is formed that positively affects yields," Egor Yershov, head of the AI Color Computational Photography group and head of the color reproduction and Synthesis sector at the A.A. Harkevich Institute for Information Transmission Problems of the Russian Academy of Sciences, told Izvestia.

In addition, digital twins can serve as a platform for digital phenotyping. This is an approach in which the results of their crossing are predicted based on models of specific plants. This method avoids the cost of physical experiments. The extensive data collected in phytotrons, as well as the flexibility of various sensor systems, create the basis for monitoring the development of these technologies, he added.

— Growing plants with specified properties is a modern trend. The prospect of these methods is to create products with increased nutritional value for target groups: athletes, cosmonauts, or people who simply monitor their health or need specialized dietary nutrition," said Pavel Pashkovsky, a leading researcher at the Laboratory of Physiological and Molecular Mechanisms of Adaptation at the Timiryazev Institute of Plant Physiology of the Russian Academy of Sciences.

Photo: Global Look Press

The key tool is high—precision control of environment parameters. In particular, using the special spectral composition of light in climate chambers, it is possible to purposefully activate specific receptors in plants, he said. This allows you to influence the biochemical composition of products in a way that is impossible in nature, where the spectrum of sunlight is constant.

For example, in the natural environment, the amount of far—red and near-blue light are mutually exclusive factors. There are no such restrictions in the chemical chamber. By combining a combination of these types of radiation, it is possible to activate receptors that are normally inactive in plants, and influence photosynthesis processes and various properties of products.

Science and Technology: Sirius showed a teledroid robot and frost-resistant radio beacons

What new products were presented at the Microelectronics 2025 Forum

— The second factor is nutrition. By combining certain mineral compounds with special light exposure before harvesting, it is possible to direct the plant's metabolism to increased production of proteins or other valuable substances. Thirdly, the addition of special strains of microorganisms to hydroponic systems also makes it possible to influence the architecture of roots and the biochemistry of plants without using genetic engineering methods," Pavel Pashkovsky specified.

Photo: Global Look Press

The difficulty lies in the fact that the listed optimal conditions (light, nutrition, microbiome) are unique for each plant species and even variety. Therefore, success is possible only in close cooperation with breeders, the scientist emphasized.

The presented technologies can play a significant role in optimizing energy supply and increasing biomass productivity. This is important when organizing long-term space expeditions in the future, as well as in the process of future exploration of other planets, said Margarita Levinsky, Scientific Secretary of the Institute of Biomedical Research of the Russian Academy of Sciences.

Since different parameters have different effects on plant species, it is necessary to select optimal conditions for each one. This is a huge job, she concluded.