Identifying LTE Connectivity Hot Spots in V ehicular Environments: A Learning Approach

Due to the increasing demand of mobile Machine- Type Communications (MTC), the interaction between MTC and human services is a recent problem for cellular communication systems like Long Term Evolution (LTE). In order to reduce the negative impact of MTC on human communication, a Learning- based Channel-Aware Transmission (L-CAT) scheme will be introduced in this paper. The algorithm bases on a learning process of cellular connectivity hot spots and is designed for non- time-critical vehicular data applications like extended Floating Car Data (xFCD) transmissions for traffic forecast systems. The results based on real-world measurements show that L-CAT leads to a much faster data transmission that correlates with a more resource efficient MTC. I. INTRODUCTION AND RELATED WORK Many mobile devices follow the routes of driving vehicles. This routes are not randomly, but follow special patterns. For example the same routes are taken by a vehicle many times: the way to work or to good friends. This fact can be used in order to predict mobility of a cellular communication device that is mounted on a car. The mobility prediction is a feasible input for traffic management systems. It can be predicted how many vehicles will pass a certain highway and if this would cause a traffic jam, some of them could be rerouted. But the mobility estimation can also be useful for communication issues. In (1) a method to improve handover quality of cellular communi- cation systems by means of a mobility forecast algorithm is shown. However, the same routes correspond also to a similar communication connectivity. This fact will be used in this paper to improve non-real-time vehicular data applications. We present Learning-based Channel-Aware Transmission (L- CAT), a decentralized communication approach that uses a learning process of cellular connectivity in vehicular environ- ments. L-CAT detects good communication locations, stores them, provides them to others users and uses them for the next drive at the same route to improve the communication efficiency. The transmission scheme can be used to provide vehicular sensor data (so called extended Floating Car Data (xFCD) (2)) very efficiently to a traffic management server. This Machine- Type Communication (MTC), which can be carried e.g. by a Long Term Evolution (LTE) network, should interfere as