Система інтелектуалізованого відеоспостереження для цілеспрямованої взаємодії з автономним мобільним роботом

Abstract

Listen

Translate article

Introduction. One of the conditions for increasing the autonomous capabilities of mobile robots is their awareness of the state of the environment in which they operate. A class of so-called indoor robots requires the timely acquisition of position data both on obstacle objects and goal interest objects for the robot actions. As a rule, the sources of data on surrounding objects are sensor devices of various types installed on the board of the robot. In modern robotics, such devices are rangefinders or scanners with various data acquisition principles and video cameras. It is images from the video cameras that contain the most complete information about the state of the mobile robot’s working room. But computer analysis and content interpretation of visual data require the development of appropriate information technologies. Successful software implementations of rather complex algorithms for visual object recognition and methods for determining these object positions relative to the room coordinate system, which were previously created as separate functional modules of the intelligent mobile robot control system, can be used to create other applied artificial intelligence systems. The formation of generalized images (models) of the environment and the states of the robot in it is important for the intelligent control system of autonomous mobile robots. Images from the robot’s onboard video camera are the most informative source of timely detection of the positions of static and dynamic objects in the robot’s working areas. Using additional information from an external video camera can not only increase the general situational awareness of an autonomous robot but also speed up the achievement of goal states, as long as the external video monitoring control system has an interface for goal-directed interaction with the mobile robot intelligent control system and conforming data structures for the representation of surrounding objects. The purpose of the article is to describe the results of the development of a control system of intellectualized video monitoring in the autonomous mobile robot working environment, as an additional external source of the object’s location data. Methods. 3D modeling of technical systems and spatial scenes, content interpretation of data, and object-oriented programming. Results. The functioning of the developed system of intellectualized video monitoring has been studied. Accuracy of the positions of static and dynamic objects obtained by this system both as a separate applied system of artificial intelligence and as an additional source of data on the presence and position of objects in the working room while interacting with an autonomous mobile robot has been estimated. Conclusions. The analysis of the functioning results of the system of intellectualized video monitoring which is intended for detecting objects and determining their positions in the room, and which is described in the article, shows that such a system can have a number of independent applications. Its most productive use is in the autonomous mobile robot working environment. Wide awareness of the autonomous mobile robot about the surrounding objects due to the interaction of two technical systems of artificial intelligence, namely, the intelligent control system of the mobile robot and the control system of the intellectualized video monitoring leads to the increase in the autonomous capabilities of the robot to perform complex tasks given by the user in a shorter time.