An Automatic Test Generation Method for Functional Coverage Improvement by Code Coverage Analysis

Coverage driven is an important technology for functional verification．However,the efficiency of coverage driven verification is greatly affected by the subjectivity of the functional coverage definitions and lengthy manual test adjustment．This paper analyzes the relationship between the line covered times and functional coverage．Besides,it also establishes a statistical probability model based on the line covered times．By using this model,a novel method,i．e．,functional coverage oriented,code coverage driven and genetic algorithm based automatic test generation （FOCDGAG） is presented．This method selects genetic algorithm to analyze the line covered times and generate new verification inputs automatically．FOCDGAG enhances functional coverage rapidly without using the information of functional coverage．Experimental results show that FOCDGAG can speed up the convergence of random test generation more than 10times and constrained random test generation more than 20 times．