- Initially, for each character, in manual mode or randomly, individual parameters are set - weight, health, stress, and speed. The mass during the evacuation remains unchanged. Other parameters may change depending on the environment and the interaction of the characters with each other. For example, a panicked person will run faster than a calm or injured person. These parameters are similar to those of a character in computer games: HP (unit of health), XP (unit of experience), and so on, - said Stanislav Torgaev, head of the research team and, head of the Department of Information Technology in the Study of Discrete Structures of the Faculty of Radiophysics.
In the evacuation, the character acts following the developed algorithm (graph) of behavior, which takes into account the various states of the character, including psychological ones: “calm”, “wounded”, “poisoned”, “dead”, as well as “stupor”, “aggression" or " willingness to help". Depending on the conditions in that the character is in, it can move from one state to another. For example, a character with a low number of health changes state to "wounded", crawls, and the speed decreases. When the health is zero, the character's state changes to "dead".
Also, characters can help each other or put out a fire, which corresponds to the "willingness to help" state. Characters in this state identify persons in need of help, approach them, and calm them down in case of stupor, and lift the fallen or give a wet cloth to those who are poisoned by smoke.
- The graph describing the behavior of the characters is a prototype, and it is necessary for testing the program. Now students of the Faculty of Psychology, under the guidance of Associate Professor Valeria Petrova, are developing a description of various psychotypes. This is the psychology of human behavior in emergencies, and later we will add them to the program, - explained Stanislav Torgaev. - Now the program includes one psychotype, that is, a character with average physical data, with a low level of aggression, who is partially ready for an emergency.
When checking a model, 30 people are evacuated from a two-story building, and everything happens quite quickly. However, when other psychotypes are added, the situation may change. The algorithms developed by the scientists will significantly increase the realism of simulating evacuation processes in fires.
The system will be able to be used by units of the Ministry of Emergency Situations at the design stage and during the inspection of public places. It will help significantly reduce the number of victims in case of emergencies, in particular, in case of fires. Also, the system can be used to teach people the basics of life safety.