Dear all,
I am simulating a stationary wireless sensor network scenario. I found out if I use Shadowing propagation model, the ad-hoc routing overhead is quite huge. In my scenario there is one sink node (destination) at one side of the network and a group of cbr/udp source node at the other side of the network. And totally 50 nodes placed in 500*500 network with grid topology. After 2000 simulation time, only 26 data packets are received at the sink node under Shadowing model but with same setting but TwoRayGround model, more than 2 thousands data packets are received. It seems it very difficult to set up the route under Shadowing model from the trace file. Is this normal? I am so surprised that Shadowing can impact the performance of ad-hoc routing protocols so badly. Since in my simulation the network is stationary, I can assume the route already set up by some omniscience. I wonder if we can use Static or Manual method to set up route for mobilenode. I tried: "$ns rtproto Static". But tcl errors are reported. Can anyone tell me how to set up the route statically or manually in wireless networks? An alternate way I think is to use TwoRayGround model for routing packets only. Does ns2 provide some commands to implement this or I need to change the C++ codes by myself? Thanks a lot. Any input is highly appreciated. Kai Bai
