Wireless Sensor Networks: Overview, Applications And Technologies

What are Wireless Sensor Networks (WSNs)?

Question:

Describe about the Wireless Sensor Networks,Design And Components?

Wireless sensors are an efficient information technology and communication tools used for monitoring and communication applications. With the emergence of mobile technologies, ‘internet of things’ etc. these wireless sensors and wireless sensor networks or WSNs have become an important part of networks and communication technologies today. 

Essentially wireless area networks are based on some ‘many- tiny’ principle. That is, there are hundreds and thousands of tiny and inexpensive devices, called sensors that helps in sensing the environment, collect data from environment, compute the data and processes the same, finally can communicate with other devices. In a wireless sensor network, there are such tiny devices that forms a network and communicate with each other through exchanging information. Hence, a wireless sensor network is also specially distributed and autonomous network of sensors. In practice, such networks are being used in monitoring different environment and its attributes. For example, it can work to monitor and collect information about sound, temperature, pressure etc. from an environment. Data is transmitted through the co- operating sensors in the network. In a WSN, usually there is a central and main location. Sensors collect data and sends to that main location.

Advancement of modern communication and networking technology has made the WSNs more efficient and applicable in diverse range of applications. Usually modern networks work in a bi- directional way. Thus, it has made the process of sending control information to the sensors in an easy way.

Surveillance operations in military was the inception point for sensor based network, but now it is being used in education to health sector everywhere. Emergence of ‘Internet of things’ has made the application of wireless sensor network beyond the horizon. It is also being used successfully in machinery, factories and industrial plants for various monitoring purposes. (Dargie & Poellabauer, 2010)

The integral part of any wireless sensor network are the nodes or the sensors. There can be thousands of nodes in a wireless sensor network.

Each of such wireless sensor or node has more than one parts. Those are,

• An internal antenna and an external antenna.
• A connection between the internal and external antennas.
• A radio transceiver having these antennas.
• A microcontroller
• An interfacing electronic circuit.
• An energy source. It can be a simple battery or some energy harvesting mechanism in some embedded form.

The size of the sensor nodes can vary greatly, there may be a small box sized sensor or some tiny one like grain of dust. The cost of the nodes also vary greatly. But cost factor is dependent on the complexity of the nodes than the sizes of the nodes. (Hailing, et al., 2005)

Depending on the size and complexity of the nodes there are other factors to consider, like energy consumption, bandwidth requirement etc. A WSN will have some topology. The topology can be a star topology or it may be some multi hop mesh topology. In practice, advanced multi hop topologies are widely used. Data propagates through such networks by flooding or routing techniques.

Sensors work on some platform designed by specific hardware and software. Building this platform efficiently determines the efficiency of the network.

Building the hardware based platform for a WSN is a challenging task. There are two main things to consider, the platform should be low cost and should have tiny sensors. There are several vendors who produces such platform. Researches are going on to develop robust and tiny platforms.

Architecture of WSNs

There should be low power consumption by the hardware part. Thus, low power consuming technologies for communication and data collections should be considered for the respective hardware solutions.

Some applications require, gateways for some LAN or WAN of these wireless sensors. Internet consists of different kinds of networks, when a WSN gets connected to the Internet it needs to get connected with different types of networks. Thus, it will require some bridging technique for communicating with other types of networks efficiently. Gateways plays the role of the bridges. These gateways also helps in storing data into some central server location. (Karl & Willig, 2007)

Again the challenge is energy consumptions. The application and the software platform used in the WSNs should consume least amount of power. As there are diverse range of applications of WSNs, so, the applications and the software interfaces depend the applications. For example, WSNs are used in hostile or remote regions. These applications are ad hoc type by nature. The protocols and algorithms used in such networks should address the following factors,

• Increased span of lifetime.
• Self-configuration mode
• Fault tolerance
• Robustness

In general maximizing the lifetime and reducing the energy consumptions are the key factors to be considered. In the ad hoc applications, self-configuration mode helps the devices to automatically conserve power, switching of the radio transmitters and receiver in idle mode etc. All these in turn helps in saving energy and maximizing the lifetime.

This is an important software part of the nodes. The operating system of the nodes is a simpler and general purpose OS. Those are similar to the embedded operating systems. There are two reasons behind taking this approach. Those are,

• There is certain aim of building and using some wireless sensor network. Thus the platform are ore oriented towards the aim rather than being a general platform.
• In any case, the platform should be low power consuming and low cost. This can be done by use of low power consuming microcontrollers and various mechanisms like virtual memory etc. These helps in building low power consuming and low cost operating system based platforms.

Some examples of OS used in WSNs are uC OS, eCOS etc. there are always some real time properties with such OS. Another example is TinyOS, this is a special kind of OS for WSNs. There are event driven programming used for this OS. Multi- threading is also used here. Other than that there are OS like, LintOS, Contiki, RIOT etc.

Online database management platforms are used for sensors and WSNs also. The owners of the sensors are needed to register to the platforms and feed data from their senso0r enabled devices to the online database for storage and future retrieval. Developers can build their applications based on such huge repository of data. Some examples are Wikisensing, Xively etc.

These platforms are useful and helps collaboration between the users. These are like huge data repositories holding data sets from diverse and different kind of sources.

The architecture of such platforms include the following components.

• Interfaces and APIs for users and online collaborators.
• The middleware tier containing the business logic for sensor data management.
• A backend storage model for data storage and retrieval. The data volumes will be bigger. Thus the storage system should be designed and developed in the way for handling such workload and requests.

Dynamics of the sensor networks is a real challenge. Eventually, with time the node of the sensor networks will fail. The reason behind such failures can be running out of power, overheating, damaged by wind, sun heat etc., there may be bugs in the software. When the nodes are installed on some fixed position, still there are chances that the quality of radio frequency range may not be suitable and the network topologies of the sensor networks will change. Such changes are sue to the changes in the environment and harder to predict.

Nodes and Platform of WSNs

Compared to the WSNs, the larger scale traditional networks may be changing of configuration of a network will need changes by a huge number of people. Because a huge number of people are there behind the network. On the contrary, in case of WSN. There may be only one single person behind more than one sensors. Thus the WSNs are supposed to be self-configuring and independent of the control given by the user. WSNs are also very adaptive to the environmental changes. These features poses several challenges in researches of the WSNs. Some of the challenges in the WSNs related researches are, (Elson & Estrin, 2004)

According to Moore’s law, the hardware part of the sensors will go smaller and cheaper with time, and the sensors will become more powerful in terms of performance. But more specifically, for WSNs the nodes are needed to be more durable, cheaper, energy efficient and smaller. No matter how smaller the nodes be, the size of the WSNs will be bigger and there will be multi-tier architecture. This multi-tier architecture based platform is also needed to be faster, cheaper, energy efficient and more capable.

The localized algorithms running on the hardware platforms are supposed to run in heterogeneous hardware setups.

In practice the networks are used today, are mostly two or three tiered architecture based networks. The highest numbered tiers are usually connected to the internet. But sensor networks are a bit different. Sensor networks can merge to the Internet, other types of traditional networks, server based computing environment etc.  But on the smallest side of the network, there will be sensor based devices like ‘motes’ etc. These ‘motes’ are similar to nodes but have minimal storage and computing capabilities.

In a typical multi-tier architecture of WSN, on the one side there are such smallest nodes, and on the other end there is usually internet, in between there will be some high performing computers. This is different than traditional multi-tier client server architecture. (Elson & Estrin, 2004)

In-network processing and routing are closely related in case of WSNs. Routing is a common issue or factor for any network and specially with larger, multi-hop networks. In a multi-hop network, any two nodes are not directly connected, rather those are connected through several other intermediate nodes. WSNs are similar to the Internet in this case. But the routing process between two cases have significant differences. In case of Internet, the goal is end to end delivery of packets in the best possible way. But in the case of a WSN, the focus in on in network processing capabilities of the network. The nodes do not only route the information, those also contribute to the processing of information.

An example of such scenario is, Directed Diffusion algorithm used in routing in WSNs. Like any other traditional routing protocol. It provides end to end delivery of information in addition to in-network processing through a data centric approach. Data is generated from the sensor nodes and can be identified by some pair of values and attributes. There are two associated concepts here, called, sink and source. Sink is the node that input in ‘interest’ to the network and requests some data. Source is a node that sends data to the sink or redirect the ‘flow’ of information to the sink according to the ‘interests’ of the sink.

Operating System for WSNs

Between a source and a sink, there will be number of intermediate nodes. The node will inspect the flow before sending it toward the sink or to the next hop. In this way, application specific WSNs are programmed. This is an additional load to the application developer to build and program the nodes in such ways. It is different than an abstraction and needs explicit programming. (Elson & Estrin, 2004)

Collaboration among the sensors is very important for a WSN. A sensor can fetch very little information about some environment at a time. Only that single piece of information may not give any useful information. But when information are collected from several similar sensors, then the big picture is created. The information about the whole environment becomes available. Thus the combined knowledge from all sensors of a WSNs is necessary. But each sensor will have different geographic locations. Automatic localization of the nodes become useful in such contexts. Also, there are different time zones depending on geographic location. When some sensor changes it location or just deployed in some WSN, then it should automatically get adjusted with the time and locality. (Mao, et al., 2007)

There are several ad hoc, automated localization schemes developed by several researchers.

Data fusion is the process of integrating different signals from different and heterogeneous sensors. There are several applications of data fusion. For example, in noise reduction, source coding, process control, source localization etc.

There have been studies to implement data fusion into WSNs. But it needs actually more than data fusion, it needs ‘centralized sensor fusion’. Centralized sensor fusion is the process of gathering geographically separated and distributed sensors that are controlled by and connected to some central processor.

On the other hand, WSNs violates several other characteristics of centralized processing system. For example, as already discussed, the nodes in a WSN have their own time and locality synchronization, own data processing capabilities.

Due to these contradictory characteristic, it is difficult to implement data fusion and related signal processing technologies in WSNs.

A larger volume of raw data is generated from a WSN. This is a continuous and streaming form of live data. In practice, there is a larger database associated to a WSN. It stores and aggregates data from various sensor nodes. There are several challenges in storing, retrieval and searching of an information in this larger database that stores live streaming of data.

Privacy and security in WSN is very important. There are several difficulties to implement and ensure security and privacy in this case. It is easy to damage or alter sensors. Even attacker can fool a sensor and the sensor will collect inaccurate or fake data from the environment. Also, stealing raw information getting generated from the sensors is easier. Implementation of security software on the nodes is difficult as those are resource and energy consuming.

Wireless sensor networks are dependent on wireless data transmission. There are chances of eavesdropping from the network. On the other hand implementation of encryption techniques will pose heavy workload on the nodes because implementation of encryption techniques for the wireless network will require in-network processing and such techniques are time and energy consuming.

Online Management Platforms for WSNs

Thus, data on the WSNs are highly insecure. It will also have privacy issues. As WSNs are installed for monitoring purposes, it collects data about individuals and share those data. These data may raise privacy questions. (Perrig, et al., 2004)

There are various and diverse range of application or wireless sensor networks in practice. Some of the significant applications are,

There are different kinds of monitoring applications of WSNs. Initially WSNs were used for monitoring in different military tasks. Some of the widely used monitoring tasks that are being done by WSNs are,

This is a common application where WSNs are used heavily. The deployment of WSN is done for some specified region from where some phenomenon is needed to be monitored. Military monitoring processes using WSN falls into this category, here WSN can be used to monitor intrusion of enemies into some region. Other than that these kind of application of WSNs used in oil or gas pipelines as some geo fencing technique.

In healthcare industry technology is being used in various ways. In terms of WSNs, there are two types of medical devices. Those are implanted devices and wearable devices. Implanted medical devices are implanted on inserted into human bodies. On the other hand, wearable devices are ‘wore’ by the human and the devices are ‘touched’ or connected to the surface of human bodies.

In healthcare, both of these devices are becoming popular and widely used for monitoring physical conditions of patients by measurement of positions, locations of the patient, blood pressure, heart beats, pulse rate etc.

The related concept is body area network, it helps in collecting information about person’s fitness, health, energy levels etc. (Lo, et al., 2005)

There are different sub areas under industrial monitoring using WSNs. These techniques and implementation are used in manufacturing plants or other heavy industrial sectors.

• WSNs are used for monitoring the status of the machineries. These WSNs are programmed to monitor the machineries based on some predefined CBM or Condition Based Maintenance mechanisms. The process is becoming popular as it reduces the cost of maintenance and helps to add more functionalities to the processes. There are various situations where, the plants cannot be monitored by humans or by wired monitoring system, these wireless sensor based monitoring networks helps to overcome such difficulties. (Xu, et al., 2004)
• There are bigger civil infrastructure that may need monitoring for getting updates about its status. There may be some geo-physical application and process of data collection over some pre-defined time frame. Data logging through infrared sensors can be useful in this context. (Krishna, et al., 2006)
• One of the widely used application of WSNs is data collection from monitoring some environment. The complexity of the work cab vary greatly. The process can be simple enough to collect data from the room environment to collecting statistical information about some complex system. An important advantage of using WSNs in data collection from environment is that these networks and sensors are capable of capturing live and streaming data.
• Water management through WSNs is an important research topic these days. It helps in quality checking and management of surface water and water from deeper levels of sea, river etc. It has significant importance to water management, water pollution management etc. Both of the animals and humans will be beneficial from this kind of application of WSNs. It is also useful for waste water management.

Environmental monitoring like earth surface monitoring is another application of WSNs. It includes monitoring processes for various pollution control, disaster management etc. Some example of such controlling are,

• Monitoring for air pollution is an area where WSNs are used. In various cities around the world this process in being used for monitoring the levels and concentration of poisonous and harmful gases in the air. One of the advantage of WSN over wired monitoring systems is the ad hoc nature and the wireless communication technologies. The mobility of the sensors help to collect data from all nearby places.
• As already said, water quality management and monitoring is an example of WSNs. In this process WSNs are used for monitoring various rivers, dams, oceans, lakes etc. along with water preserved in underground areas. It helps to create water status map irrespective of manual kind of data retrieval process. The accuracy in the map are ensured by use of the distrusted and difficult cases.
• Natural disasters like landslide can be detected by a WSN monitoring process. Before or during there will be changes in the climate like slight landslide etc. WSN can capture those details through data gathering. Thus, it is possible to detect such occurrences in advance and taking required countermeasures as soon as possible.
• Forest fire is another case that is a destroying natural and environmental event. WSN can help to detect such case in advance when there is a slight example of the same, like the phase just before starting of the forest fire. This is possible by tracking environmental metrics like humidity, temperature etc.
• Other type of natural disasters like Tsunami etc. can also be detected or predicted through monitoring by WSNs. The WSNs deployed for water management can also be used for such cases.
• WSNs are used or getting planned to use in crowdsourcing concept. For example, for finding out chances of terrorism by deploying chemical agent sensors in mobile phones of common people.

Wireless Sensor Networks cannot be classified in any group of traditional networks. These are a different kind by itself. There are several key characteristics of WSNs. The significant ones are discussed below,

1. The nodes of a WSN has constraints on power consumption or battery use or energy harvesting.
2. The nodes are resilient that is those are capable of coping up consequences of node failure.
3. The nodes have mobility feature.
4. It is easier to deploy a larger scale network through WSN.
5. The scalability of WSNs are better than traditional networks.
6. The nodes and WSN has the capability of withstanding difficult environmental situations.
7. These kind of networks are easier to use.
8. The design of a WSN has some cross layer architecture. In research, the cross layered architecture of WSNs have become an important part of study. As there were several problems with the traditional laying approach.
9. The nodes of a WSN can be heterogeneous.

There are several issues with traditional layered approach. Compared to that cross layered architecture have given several advantages. The problem with traditional approaches are,

1. In this approach, sharing of information among different layers is difficult. Thus no layer will have all information.
2. The optimization of the whole network is not guaranteed in this approach.
3. This is not adaptive to the changes in the environment.
4. There will be interferences among different kind of users, environments etc. thus there will be issues like fading, access conflicts etc. in a WSN. Traditional approach in layering cannot address this issues effectively as those are designed for wired networks.

On the contrary, cross layered approach in WSNs has the following benefits.

1. It can help in optimization of the modulation. Thus there will be improves performance of the network while transmission of data in terms of data rate, QoS or Quality of Service, energy consumption etc.
2. Very basic and inexpensive computers can be used for WSNs. A processing unit with limited memory and computing power can serve well. However the base stations needs to have some better computing power as those will act as gateways to the Internet as well. There will also be some routing options in such networks.

The complex behavior of a WSN can be simulated by agent based modeling and simulation process. This is originally based on a concept called social simulation. Some simulations like OPNET, WSNet, NS2, WSIM, NetSIM etc. can simulate some wireless sensor networks effectively.

In a centralized network mode, there is a central node that controls and monitors the network and communicate with the nodes.

Now, if this central node fails, then the whole network will shut down and collapse. Thus reliability in such network is questionable. To deal with the issue, some distributed control can be introduced in such network, without changing the other constraints of the network.

Challenges of WSNs: Dynamics of Sensor Networks and Failures

In distributed wireless sensor network, this process is adopted. The distributed control architecture in WSN are useful because of the following facts,

• The sensor nodes are very prone to failure.
• The distributed control architecture will help to collect more data in a better way.
• If the central node fails then the nodes in the distributed control will have backup of data from the network. Thus the reliability of the network will be improved.

Apart from the distributed control, the sensors in a WSN are self-organizing.

A central base station stores data collected by the sensors in a WSN. This data is stored in some numerical for. There are some standards of these data. The standards are posed by OGC or Open Geospatial Consortium.

According to those standards, the interoperability of encodings of the metadata and interfaces enables and helps the heterogeneous sensors to be integrated. Also this helps to integrate the same with the Internet. Any individual can control or monitor some WSN through some web browser.

IoT or Internet of Things is a new concept related to the Internet and WSNs. It is considered to be the future of the internet. It is a world wide web of interconnected, uniquely addressable ‘objects’, based on some communication rules and protocols. Other than computers, these ‘objects’ can be smart phones, smart devices, any RFID or Radio Frequency ID tag based device. These objects are highly dynamic and mobile. Use of WSN in the ‘Internet of Things’ can open up a new dimension of internetworking. Studies and researches are going on focusing in this topic. Still there are lots of issues and challenges to solve. (Christin, et al., 2009)

Conclusion

In this report, it has discussed about wireless sensors networks in details. After a brief introductory section there are detailed discussions on the architecture, component etc. of WSN. Then the focus has been given on several challenges in the researches on WSNs. The report also discussed the application, characteristics of WSNs in different contexts. There are discussions on related topics also like simulation of WSN, web integration, distributed WSN, Internet of things and WSN etc.

References

Christin, D. et al., 2009. Wireless sensor networks and the internet of things: selected challenges. Proceedings of the 8th GI/ITG KuVS Fachgespräch Drahtlose Sensornetze, pp. 31-34.

Dargie, W. & Poellabauer, C., 2010. Fundamentals of Wireless Sensor Networks. s.l.:John Wiley & Sons.

Elson, J. & Estrin, D., 2004. Sensor Networks: A Bridge to the Physical World. In: Wireless Sensor Networks. s.l.:Springer , pp. 3-20.

Hailing, C. L. J. et al., 2005. Overview of Wireless Sensor Networks.. Journal of Computer Research and Development , 1(021).

Karl, H. & Willig, A., 2007. Protocols and Architectures for Wireless Sensor Networks. s.l.:John Wiley & Sons.

Krishna, C. et al., 2006. Monitoring civil structures with a wireless sensor network.. Internet Computing, 10(2), pp. 26-34.

Lo, B. P., Thiemjarus, S., King, R. & Yang, G.-Z., 2005. Body sensor network–a wireless sensor platform for pervasive healthcare monitoring, s.l.: s.n.

Mao, G., Fidan, B. & Anderson, B. D., 2007. Wireless sensor network localization techniques.. Computer Networks, 51(10), pp. 2529-2553.

Perrig, A., Stankovic, J. & Wagner, D., 2004. Security in wireless sensor networks.. Communications of the ACM , 47(6), pp. 53-57.

Xu, N. et al., 2004. A wireless sensor network for structural monitoring.. Proceedings of the 2nd international conference on Embedded networked sensor systems, pp. 13-24.