THESIS
1995
ii, 83 leaves : ill. ; 30 cm
Abstract
The conventional prediction-based centralized channel assignment, which has worked reasonably well in the cellular radio networks, requires a central processor to gather information from all ports in the network to determine a frequency plan in fixed channel assignment or a channel for an incoming call in dynamic channel assignment. It requires accurate prediction of minimum channel separations as the channel assignment constraints. However, in the emerging personal communications (PCS/PCN) using microcells and picocells, the number of base stations increases so drastically that the central processor would be overloaded. In addition, the propagation environments are so diverse that it is almost impossible to "predict" a design rule for separating potential interference under all conditi...[
Read more ]
The conventional prediction-based centralized channel assignment, which has worked reasonably well in the cellular radio networks, requires a central processor to gather information from all ports in the network to determine a frequency plan in fixed channel assignment or a channel for an incoming call in dynamic channel assignment. It requires accurate prediction of minimum channel separations as the channel assignment constraints. However, in the emerging personal communications (PCS/PCN) using microcells and picocells, the number of base stations increases so drastically that the central processor would be overloaded. In addition, the propagation environments are so diverse that it is almost impossible to "predict" a design rule for separating potential interference under all conditions. These problems make the prediction-based centralized channel assignment impractical in the new networks. The measurement-based distributed channel assignment approach is easier to implement and has a greater potential to adapt to the diverse propagation environments.
In the thesis, we explore several different measurement-based distributed algorithms for both quasi-fixed frequency assignment and dynamic channel assignment under different propagation and traffic conditions. Their performance in dynamic channel assignment is also compared with one prediction-based centralized dynamic channel assignment. It was found that a simple aggressive algorithm, called Least Interference Algorithm (LIA), adapts to different propagation and traffic conditions without the use of any threshold and it outperforms other threshold-based and constraint-based algorithms.
Post a Comment