This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
SURVEY OF WIRELESS SENSOR NETWORKS.
APPLICATIONS OF SENSOR NETWORKS:
Sensor network is one which will consist of different types of sensors, some of them are seismic, magnetic, thermal, visual, infrared, acoustic,humidity and photo which are capable of monitoring a huge variety of ambient conditions such as temperature, humidity, vehicular moment, lightning condition, pressure, noise level, presence or absence of some types of objects, mechanical stress levels on attached objects, current characteristics like speed , direction and size of object.
Sensor nodes can be used for a wide range of purposes like continous sensing , event detection, event ID, location sensing and local control of actuators. We can group these applications into the following areas such as military, environment, health, home and also other commercial areas.
The various characteristics of sensor networks such as rapid deployment , fault tolerance , self organization makes them a challenging sensing system for military commands, intelligence, computing, control, surveivellance, communications, targeting(c4isrt) ,reconnaissance systems.Even though some nodes are destroyed due to critical conditions , it doesn’t effect the military operations .It will not greatly effect the military operations as much as destruction caused by traditional sensor. It strives to make the concept of sensor network a good way for battle fields.Here are some of the military applications for sensor netoworks such as monitoring friendly forces, equipment and ammunition, surveillance of battle field, reconnaissance of opposing forces and terrain, targeting, assessment of battle damage, detection and reconnaissance of nuclear, biological and chemical attack.
ENVIRONMENT APPLICATIONS: Here are some environment applications of sensor networks which include detection of forest fire, detection of flood , monitoring environmental condition, chemical or biological detection , biocomplexity mapping of environment, precision agriculture, tracking the moment of birds and small insects, irrigation , meteorological or geophysical research.
Here are some applications for sensor networks liketracking and monitoring doctors and patients inside hospitals, monitoring the moments and internal processes of insects and small animals, integrated patient monitoring, telemonitoring of human physiological data, drug administration in hospitals, providing interfaces for the disabled.
There are two types of home applications such home automation and smart environment.
COMMERCIAL APPLICATIONS: The different commercial applications are as following.
Ø Interactive museums
Ø Managing inventory controls
Ø Robot control and guidance in automatic manufacturing environments
Ø Instrumentation of semiconductors processing chambers
Ø Environmental control in office buildings
Ø Detecting and monitoring vehicle thefts
Ø Tracking and detection of vehicles
Ø Monitoring disaster area
Factors Inflencing sensor network design:
The design of a particular sensor network will be influenced by various factors like fault tolerance, production costs, scalability , operating environment, sensor network topology, hardware consriants, transmission media and power consumption. These factors are very essential for the design of a protocol.
Fault Tolerance or Reliability:
The failure of sensor nodes because of power and environment interface must not affect the overall task of sensor network. Fault tolerance is the ability to sustain functionalities without any interruption due to failures. The reliability and fault tolerance of a sensor node is modeled as the poisson distribution, to capture the probability of not having failure within the time interval
Scalability: For studying the concept the number of nodes may be a few hundred or even a few thousand based upon the application.
V(R) = (Nπ)/A
N is the number of scattered sensor nodes in the region A, R is the radio transmission range and V(R) is the number of nodes within the transmission radius of each node in region A
The price of every single node is very essential to justify the overall cost of network. Suppose if the cost of network is expensive when compared to deploying traditional sensors then it is not a cost justice. Hence the cost of sensor node must be less than $1 according to literature in order to the sensor network to be feasible.
A sensor node as shown in previous figure is made up of a sensing unit, processing unit, transreceiver unit and power unit. It can also contain some application dependent additional components like location tracking system, power generator and a mobilize. Some of the constraints associated in designing a sensor node are small size and consume extremely low power and operate in high volumetric densities and have low production cost. The life time of a sensor nework depends on lifetime of power resources of the node.
Sensor network topology:
Maintaining the sensor network topology is a challenging task due to the deployment of high density sensor nodes which are prone to frequent failures. Three phases involved in topology maintenance are predeployment and deployment phase and post deployment and Re deployment of additional nodes phase.
Sensor nodes are densely deployed inside the area of interest. They may be working in busy interactions, in the interior of large machinery, at the bottom of ocean, inside a twister, in a biologically contaminated field, in a battlefield beyond the enemy lines, in a large warehouse, attached to animals. We can observe that sensor nodes are expected to work under high pressure, extreme heat and cold and in extremely noisy environment.
For enabling the global operation of sensor networks the chosen transmission medium should be available world wide. One option for radio links is the use of industrial , scientific and medical bands which offer license free communication , free radio , huge spectrum allocation and global availability.
The lifetime of a sensor node strongly depends on battery lifetime. The power consumption of node is very important as it not possible to replace the battery once the node is deployed. Due to size factor it is not preferable to use a recharging unit like solar cell. The power consumption of sensor node in a sensor field can be divided into three domains as follows
Sensor nodes scattered in a sensor field
Ø Data processing
Sensor networks communication architecture:
The sensor nodes are deployed in the sensor field . Every sensor node collects data ,processes the data and also routes the data towards sink and end users. The sink may communicate with the task manager node through internet. The protocol stack which consists of application layer ,transport layer, data link layer, physical layer and additionally consists power management plane, mobility management plane and task management plane.
This layer will be mainly responsible for frequency selection, carrier frequency generation, signal detection, modulation , transmission and receiving techniques .
Data link layer:
The limited bandwidth and time varying nature of wireless channel, combined with radio propagation loss and broadcast nature of radio transmission, make communication over a wireless channel inherently unreliable. The data link layer is responsible for multiplexing of data streams , data frame detection, medium access and error control.
Media access control Protocols:
MAC protocols are used to create the infrastructure of network and to fairly and efficiently share the communication channel along multiple users. There are two fundamentally different ways to share the wireless channel bandwidth among different users.
1) Fixed assignment channel access methods
e.g TDMA, FDMA, CDMA etc.
2) Random access methods
e.g CSMA etc.
The MAC protocol for sensor networks should be having built in power conservation, mobility management and failure recovery strategies.
There are many responsibilities for the network layer. Some of them include routing of data supplied by transport layer and providing inter networking with external networks like sensor networks. Many routing schemes are proposed for the wireless sensor networks. Some of them are as follows.
1) Small minimum Energy communication Network.
4) Sensor protocols for information
5) Sequential assignment routing
6) Directed diffusion
In principle all of these routing protocols try to satisfy the following design criterion.
Ø High power efficiency
Ø Support for data centric
Ø Data aggregation
Ø Attribute based addressing and location awareness
Transport layer and application layer:
The transport layer helps to maintain the flow of data if the sensor networks application requires it. Depending on the sensing tasks , different types of application software can be built and used on application layer.
Power management plane:
The power management plane monitors and manages power using sensor node in the field. It may turn off the receiver during silence periods to avoid some duplicated messages. If the power level of the sensor node is low, it intimates its neighbors about its state and withdraws itself from participating in routing messages.
Mobility management plane:
The mobility management plane detects and registers the movement of sensor nodes and in turn each sensor node keeps track of its neighbor plane.
Task management plane:
The task management plane balances and schedules sensing tasks .Every management plane
should make sure that sensor nodes work together in a power efficient way so the lifetime of overall network increases.
SIMULATION OF MAC PROTOCOLS FOR WIRELESS SENSOR NETWORKS:
Medium access control protocols will be used to create predefined ways for multiple users to share channel and will enable the successful operation of network.We have fundamentally two ways to share the wireless channel bandwidth among different nodes: fixed assignment channel access methods, frequency division multiple access and Random access methods. Fixed assignment MAC protocols allocates each user an amount of bandwidth. As every node is allocated a different part of a spectrum there will be no collisions between the data whereas fixed assignment schemes are inefficient when all the nodes do not have data to send because scarce resources are allocated to nodes that are not using them. Whereas random access methods will not assign the users fixed resources. These are contention based schemes. Here the nodes that have information to transmit must try to obtain bandwidth. Some of the different characteristics and specialized applications of wireless sensor networks make the design of MAC protocols a difficult task. Some of the important features of MAC protocol for wireless sensor networks are mobility, energy, latency, throughput,fairness, collision, overhearing, control packet overhead and idle listening.
It majorly targets at reducing energy consumption through applying a periodic sleep and listening schedule. When coming to the listening mode, nodes switch on their transceiver to transmit and receive messages. When we come to the sleep mode they switch off the transceiver to conserve energy and set a local timer to make themselves wake up. Some of the features of SMAC will be described below.
Periodic listen and sleep:
The node density will be high in wireless sensor networks and if no sensing event occurs the nodes will be idle for long duration . When the two nodes are communicating with each other the neighborhood nodes will go into sleep mode until it completes the data transfer. The periodic listening and sleep patern will be used by SMAC as shown. For simplicity sake it is assumed that same schedule is followed by each node. Here we can also select different schedules by each different node if we are able to share and broadcast the schedules with all other nodes. This is supported in and improved version of SMAC in which we can adjust the duration of listening period by changing the duty cycle. Supposingly if multiple nodes want to interact with a node which is in a listening state they need to contend for the medium. The contention mechanism will be same as IEEE 802.11 which will use CTS and RTS packets. The node which sends RTS packets first will be winning the medium and the receiver replays with the CTS packets. The nodes will not follow the regular sleep schedule till the completion of entire data transmission .
To avoid collisions a same procedure of IEEE 802.11 will be followed by SMAC. In order to avoid collisions it will use both physical and virtual carrier. There will be a duration field for each transmitted packet in virtual carrier sense which will indicate the remaining transmission time. So if the node receives a packet destined to another node it knows how long it has to keep silent.The node records the value in a network allocation vector and sets timer for it. The node first looks at the network allocation vector to send data. It comes to the conclusion that the medium is busy when the value is not zero. Physical carrier sensing is performed at the physical layer by listening to channel for possible transmissions. The medium will be free if both virtual and physical carrier sense indicates it is free.
The periodic listen and sleep will decrease the time spent on ideal listening, latency is increased due to periodic sleep of each node. The delay will accumulate on each node. If each node strictly follows sleep schedule, then these will be a potential delay on each hop. So SMAC will begin a mechanism called adaptive listening where nodes are switched from low duty cycle mode to a active mode. The main principle of adaptive listening is that the node which overhears its neighbors transmission wakes up for a short period of time at the transmission end. In this approach if the node is in the next approach hop then its neighbors will immediately pass the data to it by not waiting for scheduled listened time. Suppose if the node receives nothing during the adaptive listening then it goes to sleep until its next scheduled listen time. Even though a low duty cycle SMAC is more energy efficient it even has some disadvantages. It mainly increases the packet delivery latency. Intermediate node will be waiting as long as the receiver wakes up before it can forward a packet. This is termed as sleep latency in SMAC which will greatly increase the number of hops. The traffic variations in a sensor network will not be easily adapted by the fixed duty cycle. To overcome those disadvantages DMAC was introduced.
To solve the problem of sleep latency for data gathering tree topologies the data gathering MAC is used. For a sensor network application which is having multiple sources and one sink the data delivery paths from source to sink are in tree structure which is termed as data gathering tree. For a DMAC protocol flows in the data gathering tree will be single direction from sensor nodes to sink. The main insight for designing a MAC for data gathering tree is that it is feasible to stagger the wakeup scheme and the packet flows continuously from sensor nodes to sink. In this protocol we stagger the activity schedule of nodes on the multiple half path waking like a chain reaction. The 2 figures show the data gathering tree and the staggered wakeup scheme respectively. An interval will be divided according to receiving, sending and sleeping periods. A node will be receiving a packet and sending it to ACK packet to the sender in the receiving state. A node will be turning off its transceiver to save energy. The duaration of sending and receiving periods will be the same, let it be µ. The time duration (µ) will be sufficient for the transmission and reception of a single packet. Based on the tree depth d in the data gathering tree , a node skews its wake up scheme (d*µ ) ahead from schedule of sink. In this DMAC various packets like RTS and CTS are not used since they add overheads and increase energy consumption. The sending node will be queue the packet as long as next sending slot in case no ACK packet is received. After a small number of retransmissions the packet will be dropped. The sending and receiving slot length µ is given by
µ = BP+CW+DATA+SP+ACK
BP is back off period,
CW is contention window,
DATA is packet transmission time,
SP is short period,
ACK is ACK packet transmission time.
Duty cycle adaption: If a node has multiple packets to send in a sending slot, it should increase its own duty cycle and request other nodes on the multi hop path to increase their duty cycles. A more data flag will be used to indicate request for and additional active period. The packets more data flag will be set if nodes buffer is not empty. The receiver sends ACK packet by setting more data flag when it receives a packet where more data flag is set. Suppose if a node receives a packet with more data flag then it goes to receiving state after sleeping for short duration.
Data prediction: For an instance node C has two children A and B as shown in fig 3.3 then both the children will have only one packet to send in every interval. Only one child will be able to win the channel and send the packet to node C at the sending slot. Let us assume that A wins and sends a packet to C without setting more data flag because buffer is empty. So node C goes to sleep after its sending slot without a new active period. The packet of B should be queued until the next interval. This results in sleep delay for packets from B. If a node in receiving state receives a packet it analyzes that its children still have packets for transmission. It switches back to receiving state after sleeping for a short interval after a sending slot. The remaining nodes on the path schedule an additional receiving slot accordingly on receiving this packet. In this additional slot if no packet is received the node will go to sleep directly without a sending slot.
More TO Send (DMAC -MTS): Let us assume the two nodes D and E in interference range with different parents F and G respectively as shown in 3.3. In the sending slot assume D wins the channel and transmits its packets to F. Neither E nor its parent G holds additional active slot in this interval. Hence E can only send its packet in the sending slot of the next interval which leads to sleep latency. In order to avoid this problem DMAC uses more to send (MTS), which adjusts duty cycle under the interference.
The MTS packet is very short having only destinations local ID and flag. The node sends a request MTS to its parent , if it can not send a packet as the channel is busy as well as when it receives a “request MTS” from its children. A node sends “ clear MTS” to its parent if the three conditions are true. They are as follows.
1) If the buffer is empty
2) All requests MTSs received from children are cleared
3) It had send a request MTS to its parent before and yet not send a clear MTS.
A node which sends or receives a request MTS will keep waking up periodically for every short interval. The energy consumption will be increased due to MTS packet in DMAC –MTS. Finally the use of MTS will significantly reduce latency and increase delivery ratio.
We will discuss about the models which will be used for channel propagation and communication energy dissipation.
Channel propagation Model:
In a wireless channel, the electromagnetic wave propagation can be modeled as falling off as a power law function of the distance between the transmitter and receiver. There would be no direct, line of sight path between the transmitter and receiver, the electromagnetic wave will bounce off objects in the environment and arrive at the receiver from different paths at different times. It results in multipath fading which can be roughly modeled as power law function of the distance between the transmitter and receiver . The receiver power decreases as the distance between transmitter and receiver increases. For all the simulations described in this dissertation the two ray ground propagation model is used. In this model the transmit power is attenuated according to the two ray ground propagation equation as shown below.
The received signal comes from both the directed path and a ground reflection path. Due to the destructive interference when there is more than one path through which the signal arrives, the signal is attenuated as
d is the distance between the transmitter and receiver.
Radio Energy Model:
There was a tremendous amount of research in the field of low energy radios. Various assumptions were known about the radio characteristics including energy dissipation in the transmit and receive modes which will change the advantages of different protocols. In this approach we are assuming a small model where the transmitter dissipates energy to run the radio electronics and the power amplifier and receiver dissipates energy to run the electronics. The power attenuation is dependant on the distance between transmitter and receiver to invert the attenuation by setting the power amplifier power control can be used. To transmit a K bit message to a distance d the radio will expends
And to receive this message, the radio expends:
Various parameter values of simulation set-up
Radio Transmission Range
Initial Energy of node
Figure 3.8 shows the variation of energy consumption in Joules with varying number of sources. In the second set-up we consider15 volts in which we varied source report interval and observed variation of average packet latency with varying source report interval. The figure 3.10 shows variation of average packet delay with varying source report interval as well. Further the figure 3.11 shows variation of energy consumption with varying source report interval.
From all those results we can conclude that DMAC protocol is an energy efficient protocol with low latency comparative SMAC. We can also infer that DMAC-MTS is better than DMAC-MORE. To the application which need data exchange between arbitrary sensor notes, DMAC cannot be used. Where as SMAC is a good option.
Adaptive Power Control in Mobile Environment
Wireless sensor applications work in dynamic environmental conditions and will urge for a protocol. We are going to focus on mobility analysis of both sensor MAC and Data gathering –MAC . We will also discuss a distributed power control algorithm which will adjusts to the transmit power of the node to achieve power savings in the presence of mobility and RF noise. The support of mobility
effects other characteristics like number of collisions , overhearing and delivery ratio. Battery powered wireless devices should be conserving power. The primary sources for power consumption are duration of radio transmission, power level at which the radio transmits packets and the amount of power consumed by radio when it is idle. Finding the minimum transmission power which maintains connectivity is basically simple but challenging to implement. When we think about the concept the receiver just subtracts the received signal strength from the transmitted power of the packet to obtain the current path loss and it transmits the packet to the sender. Practically noisy RF reception environment results in fluctuations in the RSS/path loss between packets although the sender and receiver are stationary. In sensor networks sink and source could be mobile leading to further fluctuations in RSS. Hence the design of practical distributed algorithm for transmit power control should adapt to node mobility and RF noise fading.
Proposed Adaptive transmit power control algorithm
The main approach is to reduce transmit power to the main level that still maintains connectivity despite intervening path loss, fading and mobility. For finding the minimum level the receiver computes the unidirectional path loss for each packet. The path loss equal to the difference between the transmitted power and the received signal strength. The optimal transmit power between a sender and receiver pair is given as
We have simulated SMAC and DMAC for 100 nodes with random placement in a 1000m×500m area under mobile environment. We have considered random mobility model and assumed inter message interval as 3 sec. Most of the specifications are similar to those in listed in table 3.1 and 3.2 ofchapter3. All the simulations are done independently under 5 seeds in NS2. Delivery ratio is chosen as a metric to evaluate the performance of these protocols under mobility. Delivery ratio is defined as the ratio of successfully delivered packet to the total packets originating from all sources. Fig 4.2 shows the variation of delivery ratio of SMAC ,DMAC –MTS and DMAC-MORE with increasing speed of nodes. If the mobility of nodes increases the delivery ratio decreases because some nodes may move out of transmission range, leading to loss of data packet. In certain cases, nodes may move closer, there by increasing interference and causing low delivery ratio. Here the decay rate of delivery ratio is not high. Fig 4.3 shows the variation of delivery ratio of SMAC ,DMAC-MTS and DMAC-MORE with the increasing speed of nodes in steps of 0.5 m/s taking initial speed of each node to be 0.5 m/s . Upto node speed of 3 m/s , delivery ratio of DMAC-MORE and DMAC-MTS is found to be almost constant whereas for SMAC it decreases linearly. After 3 m/s the decay rate of delivery ratio of SMAC is very high and the delivery ratio almost tends to zero whereas for other decay rate is linear. Fig 4.4 shows the variation of delivery ratio of DMAC –MTS and DMAC –MORE with increasing speed of nodes in steps of 1 m/s taking initial speed of each node to be 5 m/s. We had considered only DMAC-MORE and DMAC-MTS as SMAC does not support high mobility.
We had undergone simulations again by changing radio model as described in adaptive power control algorithm. Remaining specifications are considered as same as that of previous simulations . Fig 4.5 shows the variation of delivery ratio of SMAC, DMAC-MTS and DMAC –MORE with increasing speed of nodes in steps of 0.1 m/s taking initial speed of each node to be 0.1 m/s . Fig 4.6 shows the variation of delivery ratio of SMAC, DMAC-MTS and DMAC-MORE with the increasing speed of nodes in steps of 0.5 m/s, taking initial speed of each node to be 0.5 m/s . Fig 4.7 shows the variation of delivery ratio of SMAC, DMAC-MTS and DMAC-MORE with the increasing speed of nodes, in steps of 1 m/s , taking initial speed of each node to be 5m/s. The simulation results show that the rate of decrease in delivery ratio with increasing mobility of nodes is less in DMAC as compared to SMAC in both cases. In this respect we also further infer that DMAC, DMAC-MTS is better than DMAC-MORE. Our adaptive power control algorithm increased the extent of mobility support of both SMAC and DMAC by giving good delivery ratio.