|
|
I unfortunately spent most of this week unsuccessfully looking for the hardware assumptions used in the simulator. Basically I needed to find the origins of three hard-coded power values. These were the transmit, receive and idle power values of the specified device. At our meeting on Wednesday, Professor Peh decided I should simply use other values that were measured on the actual wireless card used by other members of our group for ZebraNet. In addition to this, this week I plotted the average remaining energy values among the nodes for 16 different scenarios. These cases had one of four pause times, and one of four maximum speeds. These two variables control the mobility of the nodes in the simulation. Each node moves at a random speed between 0 and the specified maximum to a destination, at which it remains for the specified pause time. I am running these simulations to find trends in the energy consumption in these different cases. We predict that when nodes move with faster speeds, they are more likely to go out of range in shorter periods of time, possibly causing less communication and therefore less energy consumption. We also believe that longer pause times reduce the number of route computations between a source and destination, but could cause nodes to remain out of range for long periods of time. Since even proving these predictions will still not suffice to come up with a power model, next week I will run more simulations in which I will measure a more detailed breakdown of the power consumption.
|
|