Modeling And Simulated Based Model Verification Of Key Components Of Mobile Robot Considering Epistemic Uncertainty

As the industry develops,the number of applications for mobile robots increases,which in turn places greater demands on the reliability of mobile robots for safe obstacle avoidance and crossing on complex roads.Due to the experimental and customised nature of mobile robots in many industries,there is a lack of data and uncertainty caused by reliance on the subjective experience of the researcher in conducting reliability analysis and simulation experiments,which affects the accuracy and credibility of reliability analysis and simulation.Therefore,it is a key issue that must be addressed to ensure the reliability of the analysis results by modelling and simulating the reliability of mobile robots under the conditions of small samples and the presence of cognitive uncertainty.The thesis is based on the analysis and simulation of the reliability of key components of mobile robots operating on complex road surfaces.The paper makes use of quantitative representations of cognitive uncertainty to model and simulate the reliability of mobile robots.The model is verified and validated to ensure the reliability of the analysis and simulation of the mobile robot.The main research elements of the thesis include:(1)Dynamic operational reliability analysis of chassis systems based on fuzzy theory and multi-state Bayesian network modelFirstly,a general-purpose mobile robot is selected for its applicability and its chassis system is structurally classified and analysed for failure.The paper then establishes a corresponding fault tree model.In order to quantify the influence of cognitive uncertainty in the reliability analysis process,the article uses the triangular fuzzy number in fuzzy theory to characterise the uncertainty in the modelling of the mobile robot chassis system.It also considers the existence of multiple failure states of the robot components and uses a multistate Bayesian network for the modelling process.The article then transforms the fault tree model into the corresponding fuzzy polymorphic Bayesian network model.The reliability analysis of the wheeled mobile robot chassis system based on the fuzzy polymorphic Bayesian network for dynamic operation in complex road surfaces is finally completed.(2)Verification and simulation of a Bayesian inference-based dynamics model for a wheeled mobile robot chassis systemThis paper completes the kinematic and dynamical analysis of a general-purpose Mecanum four-wheeled mobile robot,and then establishes the corresponding mathematical models.The uncertainties in the dynamics modelling process are quantified and evaluated for each model using Bayesian inference.The dynamics model that best fits the realistic physical model is selected by the above method,and the simulation model of the key components of the mobile robot is completed accordingly.The paper analyses the road excitation function under complex road conditions and finally completes the simulation experiment of the mobile robot chassis system operating dynamically on a complex road.It is finally verified that the chassis system can play a good damping effect on the set uneven road surface,and can guarantee the safety and reliability of the mobile robot's dynamic driving.(3)Verification and validation of a mobile robot simulation model based on interval analysis and area metricsThis paper validates the credibility of the simulation experimental results of the dynamic operation of chassis systems on complex road surfaces.To address the current situation that the model validation metric under the condition that the prediction model is a multiple output model fails to consider the coexistence of cognitive uncertainty and stochastic uncertainty,interval theory is used to characterise the cognitive uncertainty line.It is then combined with an area measure to propose a model validation metric that can measure mixed uncertainty conditions.The multiple output parameters are then reduced in dimensionality using the Marxian distance to reduce the computational effort.Finally,this paper completes the model validation between the dynamic driving simulation results of the mobile robot and the control group of experimental results,guaranteeing the credibility of the simulation results.