Critical Technologies Of Packet Deliveries For Delay Tolerant Mobile Sensor Networking

In Delay Tolearant Mobile Sensor Networking (DTMSN), all nodes cooperate to achieve a common purpose, and fairness is less of a concern. Instead, wireless nodes are mostly concerned with energy consumption and sensing applications may value low latency or high reliability over fairness. Tradeoff between energy savings and performance can be approached from three aspects:medium access control (MAC), routing and buffer management, which are assumed in the context of packet deliveries in DTMSN. Packet deliveries is a fundamental issue in DTMSN, and almost all networked applcations are based on it. The major contributions in the dissertation are as follows.1. Energy Efficient and Low-latency MAC SchemeThe problem of power consumption can be approached from two angles. One is to develop energy-efficient communication protocols that take the peculiarities of WSNs into account. The other is to identify activities in the networks that are both wasteful and unnecessary and mitigate their impact. Medium Access Control Protocol, which decides how to share the common wireless channel, allocates the limited communication resources among nodes, and a good MAC scheme can save lots of energy and reduce the number of collisions and recover from such collisions quickly.The CSMA/CA technique has the disadvantage of requiring nodes to continuously sense the channel for inactivity. This requirement results in significant energy consumption, especially when nodes do not have any packets to transfer. In order to address the challenge with respect to energy constraint, many WSNs rely on sleep/wake protocols that allow a network to selectively switch off sensor nodes or let them enter low-power sleep modes. Duty cycle operation has been introduced through the S-MAC protocol. Using this operation, the activity of a node is scheduled according to a specific amount of time, called the superframe. During this frame, a node sleeps for a specific amount of time and listens to the wireless channel for the rest of the frame. The ratio of the listen interval and the total duration of the frame is denoted as the duty cycle. During the sleep interval, the radio of the node is switched off to save energy. In the meantime, the particular node is also detached from the network.(1) Adaptive Queue Contention Queue Window based MACThe S-MAC duty cycle allows a fixed number of packet transmissions during the listen period. If a node generates (or receives) more packets than it can immediately transmit, the delay that will be experienced by the packet will increase. The delay that is incurred by the MAC protocol in transmitting a packet is generally referred to as medium access delay. This is the time spent by the sender to access the physical channel and is mostly determined by the medium access control protocol in use. Static duty cycles may result in intolerable medium access delay, which builds up the packet queue. Increased queue length may cause congestion in the network, leading to some packets being lost.Combining the use of each node queue length information in two-hop neighborhood to estimate the network traffic and modifying the IEEE802.11random backoff window algorithm to achieve a variable duty cycle operation, this thesis proposes MAC scheme with based on coordination adaptive contention queue window, which performs dynamic duty cycle operation. When there is more traffic in DTMSN, the contention window becomes larger, the chance of conflict is reduced, thus energy is saved; when there is less traffic, the contention window becomes smaller, thereby reducing medium access delay.(2) MAC with Adaptive Wakeup based on Probabilistic RoutingWhen a sensor goes into sleep mode, on the one hand it can save engery, but on the other hand it introduces medium access delay, which sensing applications wouldn’t like to accept. An ideal way would be for each next hop to wake up when the transmission of the previous hop is finished. The S-MAC protocol has been enhanced with the adaptive listening mechanism to provide multi-hop awareness. Adaptive listening does not assume knowledge of the route, nor does it try to schedule all the nodes on the route of a packet that is sent. Instead, a best effort solution is provided. Adaptive listening allows nodes that overhear a packet transfer to wake up at the end of this transfer in case they become the next hop. In DTMSN, how a node identifies whether it will be the next hop relay node after the end of the upstream transmission is very important, when the node think it could be the next hop, it performs adaptive listening; otherwise, the node continues to sleep. In this way, the approach not only saves energy to some extent, and also reduces latency. This thesis combines probabilistic routing and improved RTS/CTS methods to predict whther one node will be the next hop relay node after the end of last transmission.2. Qos Routing strategyIn DTMSN, this routing problem, that is, the task of finding a multi-hop path from a sensor node to the base station, has received immense attention from the research community. In traditional routing protocols, choosing where to forward a message is usually a simple task; the message is sent to the neighbor with the lowest cost path to the destination (usually meaning the least number of hops). Normally the message is also only sent to a single node since the reliability of paths is relatively high. However, in the DTMSN settings envisioned in the thesis, things are completely different. When a message arrives, there might not be a path to the destination so the node have to keep the message for a while and each time it encounters another node it must make a decision if it should forward the message to that node or not. It might also be sensible to forward a message to multiple nodes to increase the probability that a message is really delivered to its destination. That means many message copies are injected into DTMSN. However, too many message copies may drain each mobile node’s limited battery supply faster and result in too much contention for the restricted resources of the DTMSN, so a proper routing scheme needs a tradeoff between the number of replica messages and network performance.(1) Adaptive Threshold Routing schemeThe threshold method evolves from the epidemic routing algorithm. The difference is in that the threshold method uses a threshold value to limit the number of encounter between nodes with the same message copies. When two nodes encounter, a counter is generated specific to that message. A node with the same message copy has the same encounter. When the counter reaches the threshold, the message copy is removed from the node’s buffer. The thesis proposes adaptive threshold and energy-efficient routing scheme, which aims for the minimum amount of the smallest possible message copies necessary to obtain desirable message delivery ratio.(2) Engery Aware Routing schemePhysical and Link Layer coordinated efforts to keep energy consumption small are usually obviated by routing protocols, which put their efforts on maintaining reliability and throughput instead. The thesis introduces a scheme that minimizes the number of copies and that shifts the generation of message copies to the nodes with larger residual energy, which could be most effective in extending the lifetime of DTMSN.3. Buffer ManagmentBoth energy-efficient and low-latency MAC schemes and QoS routing strategies contribute significantly to packet deliveries’performance in DTMSN. However, owing to "store-carry-forward" mechanism and "pairwise node communication" approach and limited node resources happen in DTMSN, buffer management is, so to speak, an indispensable part of packet deliveries. After some buffer management strategies such as DF, DO and HBD are introduced, this theis puts forward a strategy for message delivery properties based buffer management, which can fully take the number of the number of packet copies in DTMSN and packet lifetime and other factors into account.