Visual Stabilization Control Based On 2D Trifocal Tensor For Nonholonomic Mobile Robot

The visual servo stabilization control problem of mobile robots has always been one of the hotspots in the field of robot control.The studies on this issue are mostly based on kinematic models rather than dynamic characteristics,and most of them adopt feedback linearization,Backstepping and other methods to design the control law.These methods have achieved better control effect in a particular situation,but there are still some deficiencies in the control process.This paper focuses on the stabilization control of visual servoing for mobile robots,mainly from the following aspects:Firstly,we reconstruct the existing experimental platform according to the needs.The experiment platform consists of wireless ISRV tracked mobile robot,video transmission system and communication system.On this basis,a global camera is added and connected to the host computer to calculate and record the running trajectory of the mobile robot in real time,and the control experiment of the servo stabilization and control of the mobile robot based on the vision is realized.Secondly,to solve the problem of visual stabilization control of monocular mobile robot,a dynamic switching sliding mode visual stabilization controller is designed.Combined the computer vision technology with the control method of the nonholonomic constrained system of mobile robot,a kind of sliding mode visual stabilization control strategy consisting in a dynamic switching control law is proposed based on 2D triple tensor and sliding mode,which is based on the consideration of kinematics model and dynamics model of mobile robot.It realize the visual stabilization of mobile robot.The simulation results show the effectiveness of the algorithm.Finally,in-depth study on the stabilization control of mobile robots based on 2D trifocal tensor is carried out.A fuzzy neural network visual stabilization control method based on 2D trifocal tensor is proposed.On the basis of sliding mode visual stabilization control strategy,based on the universal approximation characteristic of neural network,the fuzzy neural network sliding mode controller is designed by combining the sliding mode control and fuzzy neural network control,using adaptive learning algorithm to adjust the fuzzy membership function parameters and the neural network weights online.The method proposed relaxes the requirement of detailed system parameters of the mobile robot and any auxiliary controller which effectively reduces the chattering phenomenon.Finally,the effectiveness of the proposed algorithm is verified through simulations and experiment.