Hidden Markov Model based user mobility analysis in LTE network

Abstract

Models of human mobility have broad application in fields such as mobile computing, network planning and resource preservation. In L TE network, the rapid increasing number of mobile broadband users requires the availability of enhanced data services and efficient mobility management. This paper focuses on service evolved Node B (eNodeB) prediction in LTE cellular network based on human mobility, and proposes a theoretical and factual method leveraging Hidden Markov Model (HMM). The usage of HMM allows us to consider trajectory characteristics of eNodeB accessed as unobservable parameters, and also the effects of individual's historical service eNodeBs. We experiment with factual data that derives from the real-world cellular network, and analyze how different parameters impact the prediction performance using control variate method. The result shows a prediction accuracy of 53% can be achieved. These findings are very significant for the location prediction problem. And the model exhibits more merit if adopted in factual communication system. Keywords-L TE; user mobility; evolved Node B predication; Hidden Markov Mode; control variate method I.[NTRODUCTION With lower latency and higher bandwidth than its predecessor 30 Network, L TE has an expected fast-growing user base in China. The throughput and delay perfonnance are demanded to satisfy certain criteria even in the mobility scenario. Specifically, in EPS Connection Management (ECM)-CONNECTED state, the mobility of User Equipment CUE) is handled by the handover procedure [1] during which the performance of TCP and UDP is degraded due to the interference between source eNodeB and target eNodeB [2]. [n order to improve communication condition and provide service of better perfonnance, researchers have paid attention to mobility pattern and prediction of individual mobility trajectory. Since individuals usually have a precise purpose or habitual route, human mobility behavior is far from random, and is highly influenced by their historical behavior. Thus, mobility trajectory based location predication is promising to obtain high accuracy to guarantee resource preservation and better service performance. This paper addresses the development and evaluation of a model that supports the analysis and inference of a statistical pattern for next service eNodeB predication. The pattern captures the sequential relations between accessed eNodeBs in a given time period for particular individuals. In this paper, Hidden Markov Model is adopted for prediction. [n this model, we account trajectory characteristics of eNodeBs accessed as unobservable parameters, and also consider the effect of each individual's historical service eNodeBs. Based on the real world dataset of province-wide cellular network, we evaluate the prediction performance of the introduced model under different parameters with control variate method. And we analyze how different parameters impact prediction performance. The overall obtained result corresponds to an accuracy of 53%. The method used and the re...

Mobility Prediction: A Survey on State-of-the-Art Schemes and Future Applications

“…If we can foreknow the next one or even more cells, radio resource such as bandwidth can be preconfigured in the cell(s) before the user visits. In [55], the trajectory characteristics of eNodeB and a sequence of historical service eNodeBs are leveraged to predict the next service eNodeB, which will exhibit more merit for network planning and resource preservation.…”

On the Challenges of Mobility Prediction in Smart Cities

“…Standard MCs are simple and easy to implement (Zhang, Dai, 2018) but their performance, in terms of precision, are often subject to certain constraints (transition matrix's values (Amirrudin et al, 2013a), movement type (Jiang et al, 2016) …). The hidden Markov models are also efficient in terms of accuracy (Zhang, Dai, 2018) (about 53% in (Lv et al, 2014), sometimes exceeds 80% in (Qiao et al, 2015) and even greater than 90% in (Amirrudin et al, 2013b)). However, their complexity can increase with the increase of the number of hidden states and the size of the history (Lv et al, 2014), like in Ultra-Dense mobile Networks (UDNs), because of the complexity of transition matrix which considers hidden and observable states (Zhang, Dai, 2018).…”

“…Classification according to the technique used: works are categorized according to the technique used. We distinguish, among others, Markov models-based works (standard and hidden Markov chains) (Amirrudin et al, 2013b;Lv et al, 2014), Bayesien networks_based works (Dash et al, 2015), machines learning-based works (Ozturka et al, 2019), data mining based works (Soh et al, 2006), etc.…”

“…In (Samaan, Karmouch, 2005) the authors claim that most of the existing approaches assume that the user travels according to a previously known pattern with regularity. Works based on the assumption of regularity of movement offers acceptable results (between 50% and 70% precision) (Jiang et al, 2016;Lv et al, 2014…), even very satisfactory sometimes (up to 90% accuracy) (Amirrudin et al, 2013b;Liou, Huang, 2005), in terms of prediction accuracy. These works were tested even with random movement.…”

Enhanced handover mechanism using mobility prediction in wireless networks

“…One of the commonly used approaches to mitigate the frequent handover issue is mobility prediction. [3] proposed a mobility prediction model based on the hidden Markov model (HMM) to predict the next service eNodeB in the long-term evolution (LTE) cellular network. Mobility prediction-based approaches show promising solutions mainly because human mobility behaviour is far from random and is highly influenced by their historical behaviour.…”