Wireless Communication using Directional Antennas
Directional Antennas Focus the energy in a desired direction: Antenna Array Desired Communication Point Geometric representation: r
Communication Links A
Requirement #1 A B Strong Connectivity
Requirement #2 B A Hop Spanner |AB| ≤ 1: minpath(A, B) has a constant #hops.
Requirements #3 B A Minimum Radius
How can A reach B? B A
Problem Statement Given: –plane point set S, fixed angle –one -antenna per point Find: –orientations of -antennas –a minimum radius r such that the induced communication graph is a strongly connected hop spanner.
B A Case = 180 o R Euclidean MST
180 o Antennas - Clustering R BFS Traversal Select non-adjacent edges of the MST.
180 o Antennas - Clustering R Select non-adjacent edges of the MST. BFS Traversal
180 o Antenna Orientations R
Basic Observation R
R B A
180 o Antennas of Radius 2 R B A
How can A reach B now? B A
(180 o, 2)-Communication Graph R Communication Graph for: = 180 o Radius = 2 Strongly connected Hop factor = 3
What about < 180 o ? Next: ≥ 120 o
K 2 K o Antennas – 3 Point Connectivity A C B Want: Strong Connectivity Plane Coverage
120 o Antennas – 3 Point Connectivity A C B Want: Strong Connectivity Plane Coverage A C B Radius = Maximum Pairwise Distance
120 o Antennas – 3 Point Connectivity A C B Radius = Maximum Pairwise Distance ≤ 2 A C B Observation: Radius = 3 also covers the unit disk around each point
120 o Antennas – 3 Point Connectivity A C B Radius = Maximum Pairwise Distance ≤ 2 A C B Observation: Radius = 3 also covers the unit disk around each point
120 o Antennas – 3 Point Connectivity A C B Radius = Maximum Pairwise Distance ≤ 2 A C B Observation: Radius = 3 also covers the unit disk around each point
Connecting 3-Point Clusters A C B Assumption: Clusters at unit distance. X Z Y X Z Y A C B
Connecting 3-Point Clusters A C B Assumption: Clusters at unit distance. X Z Y X Z Y A C B
Connecting 3-Point Clusters A C B Assumption: Clusters at unit distance. X Z Y X Z Y A C B
120 o Antennas - Clustering R Euclidean MST(S) Partition S into clusters of ≥ 3 nodes
(120 o, 5)-Communication Graph R r = 5: Each cluster is strongly connected and covers the enclosing unit halo
(120 o, 5)-Communication Graph R r = 5: Each cluster is strongly connected and covers the enclosing unit halo B A
120 o Antennas - Summary of Results Given: –plane point set S –fixed angle ≥ 120 o There exist –orientations of -antennas Such that –radius r = 5 establishes a communication graph that is a strongly connected, 5-hop spanner. Lower bound: r = 2
K 1 K 2 K 3 ≥ 90 o : Similar Approach 4 Point Connectivity D K 4 A B C
K 1 K 2 K 3 ≥ 90 o : Similar Approach 4 Point Connectivity D K 4 A B C D A B C Communication Graph Radius = Second longest pairwise distance
≥ 90 o : Similar Approach Radius + 1: Unit halo coverage D A B C Allows us to connect clusters at unit distance
90 o Antennas - Clustering R Euclidean MST Partition S into clusters of ≥ 4 nodes
r = 7: Each cluster is strongly connected and covers the enclosing unit halo (90 o, 7)-Communication Graph R B A
90 o Antennas - Summary of Results Given: –plane point set S –fixed angle ≥ 90 o There exist –orientations of -antennas Such that –radius r = 7 establishes a communication graph that is a strongly connected, 6-hop spanner. Lower bound: r = 2
OPEN What about < 90 ? This approach does not work: Strong connectivity: –each antenna must cover at least one point Plane coverage: –some antennas cover no points Conflicting criteria! Restriction too strong!