A Machine Learning Approach for Hidden Terminal Avoidance and Throughput Enhancement in IEEE 802.11ah-based Internet of Vehicle Networks

The Internet of Vehicles (IoV) is an advancement of the traditional vehicular ad-hoc network (VANET), which is a network of many entities that facilitates real-time communication between them, including cars, pedestrians, roadways, parking lots, and city infrastructure. The hidden terminal (or node) problem, which is a major problem in IoV networks can negatively impact wireless communications and network performance. Hidden Terminal Problem (HTP) occurs when two nodes in a network cannot communicate or detect each other, leading to performance degradation. This research proposes a deep neural network based throughput optimization frame work as well as a neural network based grouping scheme for grouping nodes in IoV networks to avoid HTP. Proposed Neural Networks based grouping scheme is compared with uniform grouping (UG) scheme of 802.11ah and Received Signal Strength (RSS) scheme. IEEE 802.11ah is a standard that addresses HTP by grouping nodes based on coverage area and other criteria. IEEE 802.11ah is suitable for IoV scenarios that require more coverage and less energy consumption. The Markov chain model analyses the throughput of different grouping schemes using Monte-Carlo simulation. Both analytical and simulation results of UG Scheme and NNBG are Compared for throughput. The proposed scheme is shown to improve the throughput performance of 802.11ah based IoV networks.