A Novel Parameter Estimation Algorithm Based on GHMM for Vertical Handover

Great efforts have been driven to improve the optimal decision for network selection, which is significant for vertical handover to meet user's rapidly increased requirements. Unfortunately, the instability of the basic decisive parameters is usually ignored, resulting in a high misjudgement ratio. Therefore, this paper proposed a novel parameter estimation algorithm to bridge the observed values and the inherent characters for the decisive parameters via Gaussian Hidden Markov Model. After trained by given parameter sequences in off-line phase, the parameter pre-estimated algorithm can reveal the parameters' inherent statuses in the on-line phase. Furthermore, the pre- estimated results can be regarded as normalized inputs for network selections. The simulation results show that the proposed algorithm could reduce the misjudgement ratio of the network selection algorithms, especially exhibiting a good performance for the low speed mobile users.