Embedded Systems

Constraining Graph-based Test Case Generation by Fitness Landscaping

by Ste­fan Mueller, Jo Laufen­berg, Joachim Gel­rach, Thomas Kropf, and Oliver Bring­mann
In 2017 De­sign, Au­toma­tion Test in Eu­rope Con­fer­ence Ex­hi­bi­tion DATE – REES Work­shop, pages 52-54, 2017.

Ab­stract

This paper pre­sents an ap­proach to fuse em­bed­ded sys­tems in au­to­mo­tive or in­dus­trial ap­pli­ca­tions ap­ply­ing a com­bi­na­tion of graphs and fit­ness land­scapes for iden­ti­fi­ca­tion of sig­nif­i­cant test cases. To iden­tify these test cases graph-based space ex­plo­ration with si­mul­ta­ne­ous re­fine­ment and re­duc­tion by fit­ness land­scapes is used. Met­rics to achieve test cases with high qual­ity are gen­er­ated from spec­i­fi­ca­tion re­quire­ments. The method­ol­ogy is pre­sented in an au­to­mo­tive ex­am­ple uti­liz­ing an op­ti­cal sen­sor for lane de­tec­tion. The sen­sor is em­bed­ded in an ex­ist­ing frame­work for gen­er­a­tion of syn­thetic en­vi­ron­men­tal con­di­tions.