Modeling And Control Of Soft Robot Toward Environment Interaction

In the field of robotics,soft robots have special advantages that traditional robots do not have,such as very flexible motion ability and flexibility when interacting with the environment.These advantages of soft robots enable them to work in environments and scenarios where traditional robots are difficult to handle,such as searching and rescuing,medical surgery,and the like,with a limited,unstructured environment.But at the same time,in such kind of environment,the soft robot has nonlinear characteristics due to deformation.The traditional modeling method can not accurately describe the spatial pose of the soft robot.Control based on such modeling will inevitably lead to large errors and even control instability.Therefore,how to model and control the soft robot in the interaction with the environment plays an important role in the subsequent application of the soft robot.In this paper,research is concentrated on the modeling and control of the soft robot toward environment interaction under two circumstances.Firstly,the controllable redundancy of the soft robot is modeled,and the behavior of avoiding obstacles is completed by using such redundant degrees of freedom.Then the nonlinear differential deformation model of the tip when it comes into contact with the environment is established.Finally,a hybrid vision force controller is proposed based on the deformation model.Firstly,the kinematic model of the 8-cable-driven soft robot is established based on the segmental constant curvature hypothesis.The relationship between actuation and shape variables and the relationship between shape variables and tip pose are established and so as the two mappings between actuator space,virtual joint space,and work space.Then,using the chain law of Jacobians,the velocity relationship between the actuation and the tip is obtained.Then based on the Cossersat rod theory,deformation model of the soft robot with force applied on the tip is established.Firstly,the theoretical statics of Cosserat rod is deduced in detail.The deformation kinematics is used to represent the differential relationship between the spatial pose of rod and the deformation along arc length,and the equilibrium relationship between internal stress and external force is established by basic mechanical mechanics.The constitutive model obtains the relationship between internal stress and the deformation.According to these three parts,the governing equation of a rod is finally established.Then the force analysis and boundary condition analysis of the soft robot are carried out,and the problem to solve deformation model of soft robot under gravity and tip force is decomposed into two different boundary value problems.A threestep method is proposed to solve the deformation model of the soft robot.Finally,the validity and efficiency of the deformation model and its calculation method are verified by the comparison between the simulation results and the actual shape.Then,the controllable redundant degree of freedom of the soft robot is analyzed,and the self-motion term of the mechanical arm body is designed by using the zero space of the tip Jacobian.Using the continuous deformation characteristics of the soft robot,the intermediate marker points are used to simplify the distance calculation between the body and the obstacle.The obstacle avoidance movement of the body,and the obstacle avoidance motion strategy is designed for each intermediate point.For the tip,design both the target tracking motion strategy and the obstacle avoidance motion strategy.In the tip obstacle avoidance strategy,in order to solve the local minimum problem,the radial motion component along the radial direction of the obstacle and the tangential motion component forward to the target point are combined.Finally,according to the nature of the self-motion item,the motion values strategy under the work space is converted into the actuator control values,and the real-time obstacle avoidance position controller is obtained.Finally,the Jacobian of the tip is modified based on the deformation model and a force-visual hybrid control algorithm for the soft robot is proposed.Since the Jacobian based on the segmental constant curvature cannot correctly describe the velocity relationship between the actuator and the tip under the tip force,the deformation differential model is partial differentiated by actuator variables to obtain the corrected Jacobian.Then the visual servo and force control are designed in the respective coordinate system,and the visual servo controller is designed by the image Jacobian to obtain the velocity relationship between the image feature and the spatial feature.The equivalent stiffness is established by the equivalent series spring.The two sub control tasks for the tip motion are defined in the global coordinate system,and respectively projected into the free space and the restricted space.Then the corrected actuation of the two sub control tasks are obtained by the modified Jacobian and combined to form the force vision hybrid controller.Finally,the force-position hybrid control algorithm is verified by simulation and experiment.