A rake tree scheme divides the target HRTT into several branches to match them to groups of sensor nodes.
By distributing the remaining vertices like this, the rake tree can be equivalent to the target HRTT, whereas the remaining vertices correspond to the farthest vertices of the target HRTT.
In matching with the rake tree scheme, the root node determines a rake tree from the target HRTT (([x.sub.root], [y.sub.root]), D, [alpha], V) at first.
If this fails, a branch of the rake tree can be missing and the sensor nodes that were expected to be matched to the vertices of the branch will be matched to the vertices of other branches, incurring an unbalanced rake tree, and, eventually, an incomplete HRTT in which the boundary is not hexagonal.
It eventually may make the rake tree unbalanced and the final layout may not have the boundary of the target HRTT as determined by the root node.
On the whole, unlike general rake tree, a balanced rake tree always has the same boundary as the HRTT.
RIS contains the vertices of a rake tree that are equivalent to the target HRTT except for the root vertex.
The performance of the tree-based approach adopting an HRTT layout model with three matching schemes is evaluated using the Matlab simulator .
The coverage area of the tree-based approach adopting the rake-ID space scheme (denoted as RSpace in the graphs) is actually equal to the optimal coverage that is accomplished with HRTT. This attests that it successfully deploys the sensor nodes into the target layout.