Intelligent Auxiliary Technology Of Power Supply Design Based On Cognitive Skills Graph

The popularity of electronic devices has led to new opportunities and challenges in the switched-mode power supply industry,and with them comes a test of engineers’ own engineering skills.Objective challenges in high-performance power supply design and engineers’ cognitive dilemmas have led to inefficiencies.In the coming tide of the cognitive intelligence era,based on cognitive science theories,using artificial intelligence technology,simulating and expanding human cognition is the key to have competitiveness and viability in the era.Incorporating the constraint model of the tutor learning system and the characteristics of the sequential tracking individual behavior of the planning agent,based on the knowledge graph technology,we proposed an intelligent auxiliary technique for power supply design.However,two scientific problems needed to be solved to achieve this technology: 1.the contradiction between the need to identify and decipher errors for listing failure states and the difficulty of listing all possible error representations in complex systems.2.The contradiction between the need to master the actual state of cognitive skills of engineers and the lack of information in the actual situation to achieve personalized artificial intelligence tutoring.Based on this,we proposed an intelligent power supply design auxiliary technique based on cognitive skill graph to better assist engineers in power supply design by sequentially tracking individual behaviors and reasoning cognitive levels from individual behaviors.The main work of the paper is as follows.(1)We deeply researched cognitive science theories in the power supply domain,including constraint-based declarative knowledge,rule-based practical knowledge,and the learning from errors model analysis with the combination of these two theories.(2)Firstly,to address the problem of characterizing declarative knowledge of complex systems of power supplies,we constructed a power supply declarative knowledge graph containing stock knowledge and incremental knowledge in Neo4 j.Then,to address the problem of characterizing practical knowledge involved in power supply design,we took a typical design task scenario of power supplies as an example,adopted the declarative knowledge of the power supply knowledge graph as a constraint for defining errors,and constructed a power supply practical knowledge graph through the specialization of practical knowledge by describing the adaptation of action rules to the changing environment.The correction and improvement of practical knowledge were realized and characterized,and the practical knowledge graph of power supply was constructed.(3)For the diagnostic problem of the proficiency of the practical knowledge used by power supply engineers to solve specific problems,Evidence-based Reasoning(EBR)theory was applied to find the causes of errors.The proficiency of the engineers’ practical knowledge was obtained from their behaviors through Bayesian networks inference,which were sequentially tracked and updated in a cognitive skill graph.Finally,the effectiveness of this method was verified by a use case analysis of a power supply design intelligent assisted question and answer system to assist testers in power supply design.