Wireless Sensor Networks
Key Takeaways
How wireless networks are designed for connectivity between devices.
Protocols that are best suited for particular applications and topologies.
The components that provide the functionality for a wireless sensor network.
Wireless sensor networks are at the heart of an increasingly connected world
Networks continue to grow more robust with time. The advent of the Internet of Things (IoT) in business and other sectors is continuing to push demand for more data at higher speeds to support an increase in device infrastructure.
Wireless sensor networks are the key to arbitrating the vast landscape of devices and communication protocols; this rich engineering field encompasses software, hardware, and network design, and its role in everyday life continues to grow.
Communication Protocol and Topology
One important aspect of wireless sensor network design is to choose an appropriate communication protocol and network topology. Wireless sensor networks with a single centralized control unit typically operate using a star topology. All sensors in the network connect to a centralized control unit, and the control unit collects data from the connected sensors. This collected data is sent back to a base station or stored locally.
Common Network Topologies
Summary | Advantages | Disadvantages | Examples | |
Point-to-point | Simplest network design; a dedicated line between two devices | Can operate dynamically (connection only exists during communication) can offerer unimpeded communications between two endpoints | Constrained design by nature, lacks any expandability or flexibility | Phone call (circuit switched), LAN |
Daisy Chain | A series of consecutive point-to-point connections that can form a closed loop ring. | Ring improves pathing by cutting the travel time of the longest distance in half, additional redundancy in the case of single connection failures | Linear daisy chaining can be expanded but otherwise faces the same issues of point-to-point | SCSI, MIDI, JTAG |
Bus | All devices are connected to a central transmission line | High amount of redundancy due to decentralization | No direct device-to-device communication means failure between the node and the transmission line disconnects it from the network | CPU architecture, associated data lines |
Star | A central device is connected directly to all other network devices | Any failure (besides the central device) does not affect any other communications, also scales well | Central device represents a single point of failure for the network | Internet-capable devices connected to a router |
Mesh | Devices are routed dynamically using shortest path algorithms | Network is highly durable, allows for self-healing in the event of connection failures | Price can quickly climb due to cabling multiple paths (especially true for full mesh) | Wi-Fi networks, IoT devices |
Hybrid | Topologies can be combined to cover deficiencies or enhance benefits | Varies | Varies | Tree (star-bus hybrid): used in large LANs like hospitals, schools, offices, etc. |
While a star topology only requires a single centralized control unit to gather data from the entire network, the range of the network is limited by the overall range of the control unit. To extend the range of the network beyond the range limitation of the control unit, designers can instead opt for a point-to-point connection. In this topology, each sensor receives data from a downstream sensor and relays data back upstream to a control unit or base station.
Both of these topologies have a critical issue in that they each contain points of failure that can bring down a network. In a star topology, the entire network will go down if the control unit goes offline. A wireless sensor network with point-to-point topology will partially fail if one of the downstream sensors fail, as this will disconnect the remaining downstream sensors.
One topology that requires more complicated configuration and communication while still maintaining uptime for the entire network is a mesh topology. In this topology, a centralized control unit connects to any nearby sensors and gathers data. Other sensors that cannot connect to the control unit will connect to other nearby sensors, similar to a point-to-point topology. This provides the same advantages of a point-to-point network while spreading failure points over a larger number of nodes.
How Scalability Affects Network Selection
The scalability of the network will be primarily determined by its topology and communication protocol. A protocol like Zigbee is useful in a star network, as it operates at 2.4 GHz over 16 channels up to 100 m; a single control unit will be able to coordinate up to 64,000 nodes simultaneously. The downside is the low data transfer rate in each channel is rather low (up to 250 Kbps), so it is not the best choice for sensor networks that need to transfer image data.
Another option is the serial peripheral interface (SPI), which also operates at 2.4 GHz with up to 2 Mbps baud rate. This protocol allows connection between a control unit and up to 125 sensor nodes to 100 m in range. The larger data transfer rate allows the transmission of image data in a mesh network. An excellent option for shorter-range wireless sensor networks is to use Bluetooth, especially in automotive or indoor industrial applications.
PCB Components for Wireless Sensor Networks
Sensor node design in a wireless sensor network will need to include several components:
- Microcontrollers to gather analog data from sensors
- Wireless transceiver/antenna modules for communication protocols
- Battery power packs or a small solar module for extended service time
- The sensor or sensor array
- A memory module
Maintaining uptime requires a DC power source or battery. A solar module is very useful for periodically recharging the battery, although this increases the size of the module. The memory module is very useful, as this allows the node to store measurements in the event it loses its connection with the rest of the network.
PCBs that are part of wireless sensor networks will be mixed-signal devices. Without an integrated analog-to-digital converter (ADC), the sensor will output analog signals to the microcontroller and the microcontroller will need to send digital data to the transceiver to be modulated. This modulated digital data is then sent as an analog signal to the antenna module. Isolation of the digital and analog portions of the board will be necessary to prevent common mixed signal integrity problems, typically noise from the high-speed digital side impacting the analog.
A dual IR sensor can work through some sensor network difficulties
If designing integrated transceiver modules on a data processing board is beyond the scope of the device, there are plenty of modules that can get users started building a sensor network.
Example: The Arduino Transceiver Board
The NRF24L01 Arduino transceiver board provides wireless communication and connects directly to other Arduino microcontroller boards. These boards communicate via SPI with up to 100m range and have low power consumption. They can also be placed in standby mode when not in use to save battery power.
When using premade transceiver and microcontroller boards to run a network, the sensors used in the module will send data directly to the microcontroller board, which then sends the data upstream toward the control unit. The microcontroller board can then be programmed to implement and manage the topology necessary for the network.
Cadence Unlocks Network System Design Capabilities
A great electronics design package with an extensive components library will help layout teams design boards for sensor/transceiver modules, control units, or fully integrated modules. In a wireless sensor network, these sub-systems are critical to meeting the sharp increase in network speed, reliability, and throughput. Taking designs from the conceptual stage to schematic, layout, and production in minimal time without sacrificing performance takes many complimentary design for manufacturing (DFM) features.
Cadence’s PCB Design and Analysis tools support users with expansive simulation at the circuit and system level with PSPICE and Sigrity, respectively. Modeling data can easily be incorporated into OrCAD PCB Designer in initial layouts or revisions to expedite design turnaround.
Leading electronics providers rely on Cadence products to optimize power, space, and energy needs for a wide variety of market applications. To learn more about our innovative solutions, talk to our team of experts or subscribe to our YouTube channel.