Research on key technologies of intelligent gangue sorting robot

This paper introduces the application and research status of the intelligent gangue sorting robot. This paper points out that the intelligent gangue sorting robot is mainly based on the principle of X-ray and image identification. And the high-pressure pneumatic sorting and truss robot grasping sorting are used to separate coal and gangue. The sorting actuators are mainly truss type, parallel type and series type of intelligent gangue sorting robot. The sorting actuators have fast response speed and often separate the gangue in the form of 'pulling' and 'grasping'. In the process of belt transportation, the compatibility of different gangue sizes and the optimization of movement path need to be considered in the 'pulling' of the intelligent gangue sorting robot. And the working space of the manipulator and the bearing capacity of the robot need to be considered in the 'grasping'. This paper analyzes the key technologies such as deep learning-based coal and gangue identification, unstructured multi-constraint environment-oriented motion planning of gangue sorting manipulator, force feedback-based active compliance control of manipulator and multi-arm cooperative sorting task allocation strategy and control. These technologies are used for intelligent gangue sorting robot to effectively realize gangue sorting in complex on-site environment. This paper points out that coal and gangue identification technology based on deep learning is one of the key technologies of gangue sorting robot. It still needs further research on the efficient construction method of coal gangue data set, improving the generalization of coal gangue identification algorithm, and the real-time optimization of coal gangue identification algorithm. Combined with the demand of field application and intelligent robot development, the future research directions of intelligent gangue sorting robot are pointed out. In the complex environment on site, it is suggested to improve the robustness and adaptability of the coal gangue identification algorithm. It is suggested to develop intelligent sensing and control technology for complex environment and high-precision three-dimensional pose estimation technology for gangue. It is suggested to develop intelligent gangue picking technology of gangue picking robot based on force position hybrid control. It is suggested to research intelligent gangue sorting robot underground gangue sorting technology.