The　timely　identification　of　external　intruders　is　crucial　to　the　protection　of　concerned　regions　or　targets　against　malicious　attacks.　Multi-agent　barrier　coverage　provides　a　powerful　framework　for　the　effective　deployment　of　sensor　networks　to　monitor　baleful　intruders.　This　paper　aims　to　address　the　barrier　coverage　problem　of　robotic　sensing　networks　by　developing　a　multi-layer　coverage　formulation.　Inspired　by　gas　diffusion,　a　group　of　robotic　sensors　are　scattered　to　expand　the　coverage　territory　from　a　gathering　spot.　With　the　assistance　of　a　divide-and-conquer　scheme,　a　distributed　control　algorithm　is　proposed　to　partition　the　defence　barrier　into　multiple　curve　segments　and　integrate　it　with　intruder　monitoring.　By　adaptively　scheduling　robotic　sensors　among　multi-layer　networks,　it　contributes　to　maximizing　the　joint　detection　probability　of　sensing　networks　against　external　intruders.　Moreover,　theoretical　analysis　is　conducted　to　acquire　sufficient　conditions　for　elevating　the　detection　quality　of　a　multi-layer　sensing　network.　Finally,　numerical　simulations　and　robotic　experiments　are　carried　out　to　demonstrate　the　effectiveness　of　the　proposed　barrier　coverage　approach.