IMPROVEMENT OF ROUTING AND CONGESTION CONTROL PERFORMANCE WITH CROSS LAYER SCHEME BASED ON SIGNAL TO NOISE RATIO ON WIRELESS ADHOC NETWORKS

• Main Author: Istikmal
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/36437
• Institution: Institut Teknologi Bandung

In the fifth generation (5G) wireless communication systems, the use of a device-to-device (D2D) communication system and machine-to-machine (M2M) communication will be more intensive and massive. To meet these needs, wireless adhoc networks will play a major role in saving infrastructure and frequency spectrum. An important problem in D2D communication on adhoc networks is the existence of weaknesses in routing algorithms that still choose connection with low channel quality and increased network overhead. Moreover there is a misperception of the congestion transmission control protocol (TCP) that detects congestion based on lost data packets and timeouts. Data packets that are lost and timeouts due to high bit error rate (BER) and increased search time for communication path by routing mechanisms are misinterpreted as congestion Therefore, research is needed to improve the performance of routing and congestion control with a cross layer scheme (CLS) method that utilizes signal-to-noise ratio (SNR) as channel quality information. In this research a new algorithm has been proposed for routing mechanism with selective route (SR) method based on SNR with CLS to overcome this problem. The selection process is carried out on the establishment of communication path using the SNR Threshold metric. In the proposed method, SNR information is stored on the device in the reverse route (RR) selection process. Furthermore, in this research proposed a new algorithms and models on congestion control mechanisms, with a method of adjusting the amount of packet data transmission when congestion is detected (congestion window, CW) based on SNR with CLS. In the proposed method, the detection of lost data packets and timeouts due to high BER and the increase in routing process time are indicated in the CW model with the probability packet errors based on SNR. The algorithms and models proposed in this research are evaluated using network simulators and computer simulations. In order to approach the real results on the application of D2D communication, the evaluation conducted with various propagation models, namely Two-Ray Ground and multipath fading Rician and Rayleigh which describe SNR changes to environmental conditions. To verify the evaluation results, the proposed method is compared with the previous CLS method, the standard method and the mathematical formulation throughput model. The evaluation results of the proposed method can improve throughput performance and reduce end-to-end delay, probability of packet error, and routing overhead significantly compared to the previous method. The proposed selective route algorithm and SNR threshold metric can avoid the selection of connections with low channel quality, reduce the occurrence of path breaks and able to increase the average time of transmission path connection (path life), hence can increase throughput, decrease delay and packet error probability. The concept of SNR storage in the formation of reverse routes without adding SNR information into routing control packet is more efficient and can reduce routing overhead and avoid loss of SNR information on the network. Furthermore, the proposed congestion window adjustment method based on SNR is able to reduce misperception and improve congestion control adaptation to the environment with dynamic channel quality changes and high BER in the network, thereby significantly increasing throughput.