Development of a Multi-agent Architecture for an Object Shape Recognition System Based on Data from a Depth Sensor

The article presents the concept of an object shape recognition system for an autonomous robot based on a multi-agent neurocognitive architecture. The Microsoft Kinect for Windows controller is considered as an example of a depth sensor. The algorithms of the program for pre-processing data from the depth sensor and the multi-agent neurocognitive architecture responsible for building a model of the surrounding space, where each object is associated with a certain cloud of three-dimensional points, are presented. The proposed data processing system can be used as an element of the navigation and orientation system for autonomous mobile robots used in a real environment.