Application Development Of Inspection Robot Software System Based On Software Adaptive Technology

In order to enable the background control software of the inspection robot to perform adjustments,it can adapt to the dynamic computing environment,ensure inspection performance,and improve work efficiency and user fluency from the perspective of software operation.Based on the control theory and software engineering technology,taking the outdoor inspection robot software system as the control object,a robot software adaptive system that can observe the changes in the computing environment and adjust its own resource occupation is designed in this paper.The main work of this paper is:Firstly,for the business needs of outdoor inspection robots in substations,a background control software system for the robot is designed and written,and the analysis is abstracted to build a software adaptive application architecture.The functional characteristics of the inspection robot software system are analyzed,and the adaptive design of the inspection robot software is decomposed into three layers of observation layer-coarse-grained layer-fine-grained design,which provides framework support for robot software adaptive design.From the perspective of software operation.The working stability of the inspection robot is guaranteed.Secondly,in order to make the inspection robot software adaptive system better perceive the environment changes,and respond to possible changes in the computing environment in advance,and then adjust the resource space for the computing environment exceptions that may occur in the future,and maintain the stability of the robot with high efficiency,a computing environment observation layer that integrates trend prediction is designed in this paper.At the same time,in response to the effects of frequent jitter in the operating environment of the inspection robot software system on adaptive adjustment and prediction,a moving average observation and a prediction model based on block full connection are designed,which effectively improves the problems of frequent changes of robot adaptation targets and prediction accuracy caused by jitter.Thirdly,because inspection robots have a lot of business,it often leads to complicated software systems and large volumes.Moreover,the working state of the robot needs to be considered in the adjustment process,so it is impossible to directly establish an effective mathematical model to adjust the whole.The software system is divided into multiple functional modules,the overall resource adjustment of the software system is taken as a first-level control,and the adjustment amount is passed to each functional module in the form of resource scheduling,which is specifically adjusted by the secondary control loop within the module in this paper.Because the internal adjustment speed of each module in the secondary control is different,it will lead to local abnormalities in the overall resource adjustment.Therefore,a resource scheduling method based on collaborative control is designed in this paper to achieve effective resource control of the software system at a coarse granularity and eliminate local abnormal.Fourthly,the resource occupation and control margin characteristics of each functional module of the system is analyzed,and a multivariable decoupler for the video surveillance module with the largest resource occupation is constructed in this paper.At the same time,the internal adjustment of it with adaptive robust control is completed.Aiming at the time-varying characteristics of the software system as the controlled target,an online identification feedback decoupler is designed to increase the control accuracy in this paper.Aiming at the effect of adjustable configuration parameters as controlled variables on user experience,an error tolerance mechanism is designed in this paper to reduce the frequency of configuration parameter changes.Effectively realize the adjustment of resource occupation for each functional module.Finally,software adaptive design code is written,the adaptive management thread is attached to the existing software system in a non-intrusive programming manner,and the designed software adaptive function is verified on the robot experimental platform in this paper.