Congestion control with adaptive access class barring for LTE M2M overload using Kalman filters

Abstract

In machine to machine (M2M) communication systems based on the Third Generation Partnership Project (3GPP) Long Term Evolution (LTE), the machine type communication (MTC) devices compete in a random access channel (RACH) to access the network. An MTC device randomly chooses a preamble from a pool of preambles and transmits it during the RACH. The evolved node B (eNodeB) acknowledges the successful reception of a preamble if that preamble is transmitted by only one device. To reduce the burstiness of the connection requests in heavy traffic situations, access class barring (ACB) is proposed in the 3GPP standard. Using ACB, an MTC device postpones its request in a RACH with a probability p. In this paper, we propose a new adaptive ACB scheme for congestion control of bursty M2M traffic. The optimal value of the ACB depends on the total number of MTC devices competing in a RACH. To estimate this number, we derive a joint conditional probability distribution function (PDF) for the number of preambles selected by zero or one MTC device, conditioned on the number of MTC devices that passed the ACB check. We design a maximum likelihood estimator using this PDF. We use this estimation to dynamically adjust the ACB factor. To further improve our estimation, we use Kalman filtering based on the dynamics of the system for both known and unknown arrival distribution. In addition, for each device, we derive an approximate expression for the PDF of the delay to access the random channel. At the end, an energy consumption scheme for each MTC device is proposed while the access delay should not exceed from a maximum allowable threshold. Numerical results validate our proposed schemes and show that the total service time for the proposed method is very close to the optimal case where the information of the number of MTC devices is given.