Wireless sensor networks are a revolutionary technology for information acquisition and processing, integrating the logical information world with the objective physical world. Wireless sensor networks demonstrate vast application prospects in increasingly diverse fields.
Cloud computing is a flexible method for organizing and providing IT resources. It supports distributed storage and parallel processing, with its data processing framework handling most data locally without requiring extensive remote data transmission.
Cloud Sensor
The cloud sensor integrates measurement perception, network transmission, and cloud components. Its feature is the integration of sensing components and networking components, where the data collected by sensors is directly uploaded to the cloud platform, allowing users to access the data directly from the cloud platform, thus reducing the cost of intermediate transmission processing circuits.
Advantage 1: Simple to use; once powered on, users can directly read the sensor measurement data through a computer or mobile phone;
Advantage 2: Cloud computing enables real-time networking of cloud sensors, allowing the cloud platform to utilize real-time sensor data for big data processing, with the results of big data processing fed back to the cloud sensors. The most direct application is the online dynamic calibration of sensors, using big data to eliminate local measurement errors;
Advantage 3: Convenient for customers to quickly develop secondary applications, saving the complexity of cloud construction and reducing costs.
Cloud Computing-Based Wireless Sensor Networks
1. Architecture
The architecture of cloud computing-based wireless sensor networks is shown in the figure. A group of special nodes distributed in the wireless sensor network (WSN) area is called cloud nodes. Cloud nodes have richer resources than sensors and serve two functions. On one hand, these nodes form a cloud and are part of the cloud; on the other hand, they can communicate with sensors to collect data from adjacent sensors, also known as aggregation points (point sinks). Sensors send perceived data to a specific aggregation point (via multi-hop), and the aggregation point stores the data in the cloud. Sensors that send data to the same aggregation point form a group; to distinguish this from the concept of clusters in WSN, this group is called a partition (zone). For sensors, the entire cloud appears as a virtual aggregation node (virtual sink).
2. Partition Sensor Organization
Sensors belong to a partition based on certain rules. For example, sensors can join the partition represented by the nearest aggregation point. Sensors in each partition can be of the same type or different types, forming a local wireless sensor network that is independent of other partitions. All local wireless sensor networks are connected through the cloud to form a whole. Sensors in a partition can be organized in a flat structure or hierarchical structure.
In a multi-layer WMSN structure, the same type of sensors is organized at the same level, forming a WSN; the lower-level WSN connects to the higher-level through central nodes, and the central nodes of the higher level forward data for the lower level. At the same time, the central nodes of the higher level can also process the forwarded data, scheduling the activities of sensors at that level, such as sleeping or activating. Compared to sensors that are always activated, this working mode reduces energy consumption. However, it makes the design and maintenance of sensor software more complex. Moreover, central nodes must handle and forward large amounts of data, which will accelerate their energy consumption.
In the cloud-based architecture, another way to organize partitioned sensors is shown in the figure. Sensors of the same type each form a WSN and send their data to the aggregation point (if similar sensors cannot communicate directly, different types of sensors can act as gateways). Each type of sensor forms a logically independent but physically overlapping WSN, which connects to the same aggregation point. Data processing software is deployed in the cloud (not just at the aggregation point connected to these WSNs). After collecting and processing various perception data, the cloud can (through the aggregation point) schedule the sensors in each independent WSN to operate in a reasonable manner. Compared to the cluster-based single-layer structure, this organizational method achieves smaller scales of independent WSNs in the partitions, higher transmission efficiency, and control information still has an overall network perspective.
3. Cloud Organization
When cloud nodes form a cloud, the following issues need to be considered: a) how to deploy cloud nodes in the WSN area; b) how cloud nodes communicate with each other; c) how nodes are organized into a cloud; d) how the data collected by aggregation points is stored in the cloud and processed in the cloud.
The selection of the location of cloud nodes (aggregation points) depends on the distribution of sensors in the wireless sensor network. Ideally, the location of aggregation points can minimize data transmission operations throughout the network, and complex algorithms can be used for location selection. However, due to the dynamic characteristics of WSN (topological changes, routing changes, diverse data generation locations and patterns, etc.), achieving the ideal goal is difficult. In the case of uniformly distributed sensors, a simple method is to have each aggregation point responsible for collecting data from approximately equal numbers of sensors.
Cloud nodes must communicate with each other to form a cloud. They need to be reasonably deployed in the WSN area, working in a manageable environment, so network infrastructure may exist. If a wired network is present, nodes can communicate securely using methods like VPN; if there is a mobile communication network (such as 2.5G, 3G, LTE, etc.), cloud nodes can also interconnect using these networks. However, wireless sensor networks are usually deployed in environments without infrastructure, making self-organizing wireless networks the only available communication method. For example, cloud nodes can be configured with the IEEE 802.11 protocol stack to form an Ad hoc network.
Sensor Data Upload to the Cloud
Wireless standards have undergone tremendous changes, allowing sensors to synchronize large amounts of data to the cloud and retrieve it as needed. People have recognized the potential benefits of the Internet of Things. We will see a variety of solutions, but ultimately, only those that are more useful and cost-effective will survive.
Through wireless connections, we can access cloud space on multiple devices. Below are several ways to synchronize sensor data to the cloud.
1. Wired Connection
This is the simplest method, originating in the 1970s and 1980s, and it is the ancestor of all wireless connection methods. Sensors are equipped with a microprocessor responsible for processing the collected data, which is then uploaded via a wired network. Additionally, the processor can modify or update certain functions of the sensor. However, this method has significant limitations, as it is not feasible to have wired connections everywhere.
2. Mobile Network Connection
The mobile network followed the wired network and saw initial development in the early 1980s. This network naturally became the first widely used wireless network. However, it has its downsides: first, to connect to a mobile base station, sensors must still connect to a mobile phone via a wired connection or embed a dedicated baseband in the phone (which is costly); second, the uplink wireless transmitter requires significant power support; lastly, data charges can be painful.
3. Remote Wireless Network Connection
As early as 1947, regulators opened many unlicensed wireless frequency bands, but at that time, they went largely unnoticed until the 1990s when mobile phones emerged, and people began to realize their value. In 2003, the 902-928 MHz and 2400-2483 MHz bands became popular, used for the latest IEEE 802.15.4 wireless standard.
The mesh networks mentioned earlier also use these frequency bands, consisting of numerous small, low-power wireless devices that are highly interconnected, gathering sensor data from edge areas to a collection point, all connected to the cloud. This ensures wireless network coverage.
4. Wireless Router Connection
The well-known 802.11 Wi-Fi standard was born in 1997, using the 2400-2483 MHz and 5130-5835 MHz bands. These two bands have profoundly impacted ordinary people’s wireless lives, with such routers ubiquitous in homes, companies, and public places.
Additionally, there are some specialized routers mainly used in industrial and infrastructure fields. Once a wired connection is made, the router can connect to the cloud; in fact, wireless routers are the primary means for people to enjoy cloud services in daily life.
With the emergence of Wi-Fi functionality on smartphones, sensors that can connect directly to routers have also been developed. This means that as long as the sensor is within the router’s signal coverage, it can connect to the internet at any time, eliminating the complex process of connecting to a mobile base station.
5. Mobile Phone Connection
In real life, there are many instances where sensors do not need to connect to a wireless router; a mobile phone can handle everything. This way, users can interact directly with the sensors to obtain the information they need. Moreover, in many application scenarios, there is no need to transmit data across oceans; wireless headsets are one such example.
These data transmissions can be accomplished via Bluetooth, which, like Wi-Fi, operates in the 2400-2483 MHz band. This standard was born in 1998 and became part of the 802.15.4 standard in 2003, but it continues to play its role today.
Recently, technicians have introduced a new low-power Bluetooth technology (BLE), which is energy-efficient and well-suited for low-rate or low-duty-cycle simple sensors. This technology provides strong momentum for the development of small sensors, which previously required wireless networks or mobile networks (such as various smart bracelets) and can now interact directly with mobile phones.
Recently, technicians have also added Bluetooth functionality to traditional Wi-Fi routers. This allows sensors equipped with BLE technology to connect directly to the cloud through routers, eliminating the need to connect via mobile phones.
With the continuous growth of various wireless standards, we can use more methods to promote the development of the Internet of Things.
With the development of the Internet of Things, sensor networks are integrating cloud computing technology, elevating the data processing technology of sensors to a new level. The information obtained will greatly increase, and the integration of multiple functions is an inevitable trend, which will also indirectly promote the development of sensor technology.
Disclaimer: This article is an original piece from the Sensor Technology Platform. For reprinting this article, please contact us via phone or WeChat! Reprinting without permission is prohibited! 17301083560
Note: 1. Public account name and ID 2. Title of the article you wish to reprint
