Research On Rongcha Grasping Of Soft Robot Gripper

The end-effector capability of a robot is one of the key basic functions of robots.The actuator should own the ability to complete grasping and assembling to meet common end-execution task requirements.Since the 1950s,robot grasping technology has been widely used in closed scenarios such as automobile assembly,machining,handling and palletizing.However,the current technology still has huge challenges in non-closed scenarios such as household services,agricultural picking,express sorting,and 3C assembly.The key reason for the above situation is the limitation of the existing robot end grasping mechanism:the robot decision module must calculate the precise description of the desired action based on the precise perception/model information,and then the grasping actuator executes it precisely.This leads to the inevitable need for closed transformation of the scene in the practical application of robotics,whose purpose is to ensure the certainty of the scene where the robot is located.That is,in this scenario,the robot’s workflow,operating objects are pre-defined and can be accurately observed and analyzed.Eliminating or alleviating this limitation will further promote the application and development of robotics in a large number of labor-intensive industries that are not suitable for the complete closed transformation of the environment,and effectively promote the intelligent upgrading of related industries.In a non-closed scenario,the basic assumptions faced by the robot are different from those in a closed scenario.Robot usually cannot obtain accurate environment models and precise information of operating objects in non-closed scenarios.Therefore,the grasping task in the non-closed scene mainly faces three challenges:1.The prior information such as the material of the grasped object is missing;2.The accurate object model is difficult to obtain based on the sensor information;3.Based on traditional precise control method to design grasping execution mechanism is expensive.In response to the above challenges,this paper studies the Rongcha grasping of the soft robot gripper from the three aspects of control method,execution mechanism design,and system implementation,and explores the grasping method in non-enclosed scenarios.Aiming at the problem that the grasping execution mechanism cannot obtain prior information such as object material in non-closed scenarios,this paper proposes a Rongcha control method for soft robot gripper.For the grasping task,the definition of the Rongcha grasping method is proposed,and its operational principle and working pattern are explored.The Rongcha actuator does not need to know the specific parameters of the object when facing a large variety of grasping tasks,but only needs to know a boundary conditions for large classes of objects.On this basis,this paper designs a soft gripper.Facing objects of different sizes and materials,the same control input can be used to ensure that they are grasped without damage.Embedding suction cups on the gripper can achieve "half-enveloping" grasping method with suction force,which utilizes the characteristics of flexible gripper that tends to be deformed by force.The grasping size range and load capacity of the gripper are improved through"half-enveloping" grasping method.Aiming at the problem that the grasping execution mechanism cannot build accurate object model based on sensor information in the non-closed scene,this paper proposes a grasping area selection method that does not rely on accurate object models,a strategy to adjust the hand position during grasping,and a multi-object environment grasping planning by combining object point cloud information and color images.All grasping strategies do not rely on accurate model information,which can reduce the requirement of sensor modeling.Aiming at the problem that the high cost of grasping execution mechanism designed based on the traditional precise control method in the non-closed scene,this paper proposes a integrated design method combining the design of the sfot robot gripper structure and the grasping strategy,and the low-cost Rongcha robotic grasping system is developed by using this method.In this paper,the mechanism that the end effector is deformed by force after contacting the surface of the object replaces the active feedback mechanism based on vision and decision-making to deal with the uncertainty of the shape and material of the object,which reduces the design,deployment and execution costs of the robot grasping execution mechanism in non-closed scenarios.In this paper,a low-cost Rongcha grasping system is constructed by using low-cost perception equipment in the actual scene,and the working effect of the Rongcha grasping method is tested through experiments.