This　paper　proposes　an　augmented　transition　network　(ATN)　based　algorithm　to　deal　with　storehouse　robot　navigation.　At　first,　two　operations　of　edge　detection　and　binarization　of　original　image　are　employed.　The　lane　line　equations　are　evaluated　by　scanning　the　lane　line　pixels　through　augmented　transition　network　(ATN)　algorithm　in　the　binary　image.　Then　the　angle　offset　and　horizontal　distance　offset　of　robot　can　be　calculated　from　the　lane　line　equations　and　the　art　of　geometry.　Finally　an　expert　control　algorithm　is　used　to　control　robot　navigating　according　to　the　angle　offset　and　horizontal　distance　offset.　Experimental　results　show　substantial　improvement　in　efficiency　and　reliability　with　this　algorithm.　In　addition,　the　runtime　can　satisfy　the　requirement　of　realtime　navigation　while　the　performance　is　significant　enhanced.　©　2015　IEEE.