Ad Hoc Network Performance Improvement

Due to recent increase of the demand for mobile communication, the technology of mobile communication is developing rapidly and used wildly. Cellular mobile communication system, as an example, has finished its evolution from the first genernation to the second and second and a half generation, and it is now matching forward to its third generation system.Wireless Ad hoc network is an acentric wireless network with characters of self-organized and multi-hop. Generally speaking, wireless channel is characterized by the high Bit Error Rate, which greatly limits the performance of the wireless network. In a traditional solution, an adaptive end-to-end Forward Error Correction (FEC) technology is used to reduce the loss probability of packets and thus improve the performance, but this still can not adapt the link diversity within the network efficiently. However, the concept of active network introduces a new method to implement the adaptive FEC in a hop-by-hop way, which is supposed to improve the performance of wireless link greatly. When implement the FEC technology in a hop-by-hop way, the work of coding and decoding, error detection and correction can be done in the intermediate nodes, as well as the source and destination nodes. Furthur more, the intermediate nodes could change the FEC level according to their local link condition, which can maximize the performance of the network in different link conditions.We propose two adaptive FEC algorithms based on the concept of active network. Both of the two algorithms would monitor the Packet Error Rate (PER) of the wireless links, using computation function of active nodes, and adjust the FEC level according to the detected PER level. Thus we can increase the efficiency of the utilization of bandwidth while keep the PER in a low situation. Our active network based adaptive FEC algorithms are implemented first in the execution environment ANTS using Java coding, and we simulate them under the Gilbert Bit Error Model. Then we transfer the simulation to the NS2 platform, to implement them in an Ad Hoc wireless network. The results show that both of our two algorithms can effectively reduce the loss probability and transport delay.With new wireless network devices/systems come into existence, the deployment of Ad Hoc network has changed greatly. Sparse Ad Hoc network has received more and more attention, even researchers from different fields are trying to introduce the sparse Ad Hoc network technology into their own fields. In this paper we also describe the application background and technology architecture of the sparse Ad Hoc network, we will analyze three typical sparse Ad Hoc routing protocols (Epidemic, PROPHET and Message Ferrying). In sparse Ad Hoc our goal is to decrease the transport delay, which is supposed to be a key performance index of the routing protocols here.We also propose MFBR (Message Ferry Based Routing) algorithm, in which several independent and cooperating ferry nodes act as relay nodes to decrease transport delay and increase network stability and security. According to the simulation result, MFBR routing algorithm could effectively decrease the average delay of sparse Ad Hoc network, thus improve the network performance.