Research On Automated Regression Testing Method For Embedded Software

With the rapid development of computer hardware and software technology, the scale and performance of the embedded system hardware have been greatly improved. The complexity and scale of system software and application software for embedded system both grow accordingly. Being a key stage in embedded software development process, it also provides guarantee for their quality. Embedded software testing is the important links in the embedded software development, and important means of keeping embedded software quality. It's not possible to develop a software at a time. Instead, hundreds to thousands modifications and testing are needed. All software development is impossible a one-time job. However, the revision of the software for embedded system is a time consuming work since their special complexity. Aiming to address this problem, the automated regression testing method for embedded system emerged. Now, it has been an unnegligible part in the life cycle of software development process. Research on automated regression testing method for embedded software is still in its initial stage and it's an urgent problem to find a method which can locate the bugs in the programs efficiently.This paper introduces three methods of fault localization in regression testing as follows:the method based on program chopping; the one based on program spectra and the one based on code coverage; In addition, we also introduce an object oriented method, i.e., the atom dependence analysis. Then, we propose a frequency spectrum and atom dependency based fault localization method which can be adopted to embedded system software. Detailed examples are used to illustrate our opinion. Experiment results show that our method can locate the mistakes in the software testing efficiently and effectively and the value to adopt this method in real world. At last, we introduce the process of embedded software testing and propose a new framework for it.