Spatial-Temporal Database Based Asynchronous Operation Approach of Fruit-Harvesting Robots

Autonomous fruit-harvesting robots encounter difficulties of low fruit recognition rate and picking efficiency due to the complex unstructured operational environment. To solve this problem, an asynchronous approach has been proposed to discriminate the recognition and manipulation process. The fruit recognition task can be intensified via repetitious inspection or human-robot interaction, meanwhile a spatial-temporal database is constructed to record the recognition information which might facilitate the sequential picking manipulation. In this paper the attributes of a spatial-temporal object are firstly investigated with four elementary constituents attached. Hereby the fruit target is modeled for harvest decision-making. Secondly a three layer database management system is designed as per the modular design principles. Finally, we introduced a picking scheduling application based on this database management system. The picking schedule demonstrates that the Construction of the spatialtemporal database paves the way for the success of paradigm shift from synchronous to asynchronous manipulations of fruit-harvesting robots.