Reliability Analysis Of AUV Reconfigurable Software Based On Petri Net

In recent years,with the increasing importance of human exploration of the ocean,autonomous underwater vehicle has been widely used in marine resources exploration,environmental monitoring,and underwater maintenance operations.The Complex marine environments and unknown external factors require underwater vehicle to have more robust control structures and more sensitive self-state awareness systems.And the reliability of the control system which is the core unit has become one of the most important indexes of underwater vehicles,which would affect the quality of work and even affect the survivability of underwater vehicles directly.Therefore,the study of reliability evaluation methods for underwater robot software systems has become an important issue in the development of underwater robot technology.In view of the characteristics of "poor information and less data" in the early design stage of complex component-based software system,this paper uses stochastic Petri net technology to research and implement a reliability evaluation model based on software architecture.First of all,a comparative study of existing software reliability models,the reliability white box model based on software architecture is more suitable for the reliability assessment of the initial design of autonomous underwater vehicle control software system.However,the existing architecture models generally have problems such as low parameter accuracy,excessive model assumptions,and no intuitive description of the software architecture.Therefore,this paper introduces the stochastic Petri net to model and analyze the software architecture.Next,the basic principle of the performance analysis of the system using stochastic Petri nets is studied in detail,and the calculation method for the probability of the distribution of the instantaneous state and the probability of the distribution of the steady state are obtained.Based on the classic Littlewood model,the influence of serial and parallel coupling of components on software reliability is considered.An improved Littlewood model based on stochastic Petri nets is proposed and verified by a simple system.Then,in order to alleviate the problem of model state space explosion,an improved Littlewood model based on generalized stochastic Petri nets is proposed.The idea of calculating the distribution probability of the steady state in the generalized stochastic Petri net model is studied,and the steady state probability distribution and the system performance index are obtained by using a solution algorithm based on a stochastic discrete finite-state Markov process for a simple system.It proposes a "transition fusion" method that eliminates the disappearance state,solves the problem of the use probability of transient transition,and uses this model to solve the overall reliability of the system.Eventually,by studying the types of underwater robots and using Petri nets,the control process of AUV to complete the marine environmental exploration is described,which reflects the reconfigurable features of the AUV control software system.Based on this,a hierarchical hierarchical structure of "ZT-AUV" is proposed and a generalized stochastic Petri net model is established.Then,based on the embedded Markov chain analysis,the steady-state distribution probability,module participation,and interface call frequency are obtained.The overall reliability of the software system is calculated by using software steady state operation and state transfer reliability weighted calculation.At the same time,according to various performance indicators can be targeted to improve the reliability of the software system,verify the applicability and effectiveness of the model.