Advertisement
Articles
Advertisement

Five Things Every Scientist and Engineer Should Know about Wireless Sensor Measurements

Mon, 03/09/2009 - 12:35pm
Charlie Stiernberg, Product Manager, Remote Data Acquisition National Instruments

Wireless is everywhere. Freedom from complex cabled infrastructure and restrictive wiring schemes has left no industry untouched or uninterested, including PC-based measurement and automation. Convenience, flexibility and cost reduction are some of the many potential benefits that make this technology an attractive option for data acquisition. In addition, new challenges, such as aging national highway infrastructure and ailing environment health, are compelling scientists and engineers to take measurements out of the lab and into the field. Wireless has the potential to bring measurements to areas where they have previously been impossible or impractical.

Yet, with the rapid proliferation of so many different standards and protocols, it is difficult to keep your options straight. How do you choose between ZigBee, Wireless HART, Bluetooth, Wi-Fi or any of the other myriad wireless solutions? What tradeoffs should you evaluate between different wireless technologies? Consider each of the following five key factors – bandwidth, power, topology, standardization and software – before implementing a wireless measurement system in your next project.

Bandwidth

Not all sensors are made alike. Some require excitation or bridge completion; others need cold junction compensation or low-pass filters. Similarly, not all wireless sensor measurements have the same requirements, and not all wireless solutions have the same capabilities. When designing any measurement system, wireless or wired, two important specifications to consider are bandwidth and latency.

Bandwidth, in the context of wireless communication, is the rate at which data can be sent from a transmitter to a receiver, typically specified in kilobits per second (kbps) or megabits per second (Mbps). Latency, a related specification, is a measure of the delay between transmission and reception. There are tradeoffs between bandwidth and latency for any instrumentation bus, whether you are using USB, PCI, PXI or wireless. Typically, an internal bus such as PCI has the shortest latency, because data has a finite distance to travel over short wire traces. Wireless, on the other hand, has the longest latency, because data must travel over varying distances through a shared medium: air. This is true of all wireless protocols, because it is inherent to the underlying technology. Yet, while little has been done to address wireless latency, bandwidth has continued to be an area of innovation and differentiation. If you need proof, consider all the 3G advertisements for mobile phone providers.

Wireless sensor technology has evolved over the last decade from simple, low bandwidth 900 MHz radios, to more advanced standards, including IEEE 802.15.4 and IEEE 802.11.
Figure 1. Wireless standards and protocols differ in their tradeoffs between bandwidth and power consumption.
Click to enlarge

Over the last several years, wireless sensor technology has evolved from simple, low-bandwidth 900 MHz radios, similar to those found in cordless phones and Bluetooth radios to more advanced standards, including IEEE 802.15.4 and IEEE 802.11. In addition, 900 MHz technology is popular because it is inexpensive and readily available. With a bandwidth maxing out at around 150 to 200 kbps, it is appropriate for measuring slowly varying phenomena such as temperature. Bluetooth has been a popular option for slightly faster measurements, because the volume of Bluetooth-enabled mobile phones and computer peripherals has driven down the cost of the radio chips. The IEEE 802.15.4 standard was the first targeted specifically at sensor measurements, and several organizations, such as the ZigBee Alliance and the HART Foundation, have adopted the technology. Though IEEE 802.15.4 bandwidth lies somewhere between that of 900 MHz radios and Bluetooth, many wireless sensors also include some level of onboard processing for filtering, fast Fourier transforms (FFTs) or other data reduction. Finally, IEEE 802.11, commonly known as Wi-Fi, provides the highest bandwidth for wireless sensor measurements. IEEE 802.11g is currently the most popular standard with a theoretical bandwidth of 54 Mbps, but IEEE 802.11n promises to increase that by up to an order of magnitude. Applications with dynamic signals such as vibration, sound or strain require that extra bandwidth to transmit complete waveform data. Refer to Figure 1 for an overview of these wireless standards and their respective bandwidth capabilities.

Power

Recent advancements in wireless bandwidth have been closely followed by advancements in low-power design. An increased focus on power consumption in the general PC market has translated to an intense focus in the wireless sensor sector, where battery life can be critical for some applications. Most wireless sensors, particularly those based on IEEE 802.15.4 or 900 MHz radios, are designed to use power very conservatively. In fact, many can last several years on batteries by restricting data transmissions to every few minutes, hours, or even days, and “sleeping” in between. For applications without ready access to a fixed power supply, such as outdoor environmental monitoring, these low-power devices are a necessity.

Battery longevity, however, comes with a price; bandwidth, transmission range and even measurements can take a hit. With infrequent transmissions, bandwidth becomes almost irrelevant, as very little data has to be sent back to a host application. Also, transmission power of the radio, which is directly related to the transmission range, affects power consumption of the overall system. In addition, some measurement types may only be possible with restricted use, such as those requiring current or voltage excitation.

Wi-Fi-based devices consume the most power, because the technology was designed for use with standard PC architectures. Battery operation is possible, but without a renewable source of power, such as a solar panel, the practical lifetime is limited to about a day of continuous transmission. New research in low-power IEEE 802.11 will continue to be an interesting possibility over the next several years.

Topology

900 MHz radio waves can propagate farther in free space than the 2.4 GHz waves used with Wi-Fi, Bluetooth, and IEEE 802.15.4 devices; however, 2.4 GHz waves handle obstructions better.
Figure 2. Wireless sensor networks adhere to one of three network topologies.
Click to enlarge

The need for wireless within a measurement system is often driven by the limitations of its physical environment (www.ni.com/wireless/remote_monitoring.htm). Sometimes, wireless is necessary for long-distance measurements, such as monitoring an oil field, and sometimes it is merely to overcome a physical obstacle, such as a rotating piece of equipment. Transmission range and network topology go hand in hand when designing a wireless measurement system.

Generally speaking, the larger the distance between a transmitter and a receiver, the lower the effective bandwidth and data rate. The radio frequency at which a wireless sensor transmits data also limits the transmission range. 900 MHz radio waves can propagate farther in free space than the 2.4 GHz waves used with Wi-Fi, Bluetooth and IEEE 802.15.4 devices; however, 2.4 GHz waves handle obstructions better. For some applications, extending the range of a wireless signal is as simple as using a higher gain antenna. Omni-directional antennas are the most common, available with nearly every off-the-shelf wireless router or access point. They radiate energy equally in all directions within a two-dimensional plane. Directional antennas, such as a Yagi antenna, radiate energy in one direction with a high gain. Directional antennas are good for extending the reach of your wireless system over great linear distances. To cover a larger geographical area, new network topologies may be appropriate.

Wireless sensor networks are typically organized in one of three types of network topologies: star, cluster tree or mesh. Refer to Figure 2.

In a star network, each sensor node connects directly to a gateway, which in turn provides the data to a host computer. Almost all Wi-Fi networks use this topology, where each client (sensor node) is tied to a wireless access point (the gateway). In a cluster tree network, each node connects to a node higher in the tree until the data reaches the gateway. This simple algorithm may be used with any wireless radio technology. Finally, in a mesh network, nodes can connect to multiple other nodes in the system and pass data through the most reliable path available. If one link goes down, affected nodes can pass data through another path to reach the gateway. Most IEEE 802.15.4-based devices employ some type of mesh network topology, because of this self-healing effect.

Standardization

The majority of this article discusses wireless sensor measurements within the context of four leading technologies: 900 MHz, IEEE 802.15.4, Bluetooth and IEEE 802.11 (Wi-Fi). The reality is that there is a myriad of wireless protocols, too many to provide an exhaustive list. The industry has been slow to adopt a standard for wireless measurements. Even solutions based on IEEE 802.15.4, which only specifies the physical and media access control (MAC) layers of the OSI model, often use incompatible protocols on top of similar radio technology. For example, a ZigBee device may or may not communicate with a Wireless HART device. There are arguments from both sides of this issue – those who are in favor of standardization and those who are not.

A significant amount of network infrastructure, including access points, routers, switches, and PCs, already exists at most facilities to support wireless measurement applications.
Figure 3. A survey of wireless technologies for sensor measurements reveals strengths and weaknesses for each solution.
Click to enlarge

A proprietary wireless protocol is typically developed by a vendor or end user to satisfy the requirements of a very specific application. More often than not, these protocols are built on top of a 900 MHz radio with custom data encryption, packet sizes and routing algorithms. The advantages of this approach include bandwidth, transmission range and power consumption tailored to the special circumstances of an application. For example, a temperature sensor mounted on a rotating machine may harvest energy from the vibrations of the machine to power a very simple radio. Some proponents of custom wireless solutions also argue that they are inherently more secure because the security algorithms used to protect data are themselves a secret. This is true, provided the security algorithms remain a secret; but once they are not, all future wireless transmissions (and potentially recorded past transmissions) are compromised.

Wireless sensor solutions based on standardized hardware technology and communications protocols also offer many advantages. Choosing standardized hardware and software protects your investment from the fate of a single vendor, giving your application both flexibility and longevity. Wi-Fi is perhaps the best example of a wireless communication standard, having been in use by the IT sector for more than 10 years now. A significant amount of network infrastructure, including access points, routers, switches and PCs, already exists at most facilities to support wireless measurement applications. In addition, the IEEE 802.11i security standard provides robust protection for data with 128-bit AES encryption and IEEE 802.1X port-based authentication. NI Wi-Fi data acquisition (DAQ) devices (www.ni.com/wifi) are an example of wireless sensing hardware that support these IT standards. Refer to Figure 3 for an overview of the pros and cons between different wireless technologies.

Software

Wireless hardware by itself is only half a solution, regardless of the technology in use. Without software on the receiving end of your data, your measurements are meaningless. When considering software for any measurement system, wired or wireless, there are two options: turn-key or programmable.

Using an open software platform capable of communicating with all wireless protocols is a prudent choice. Many vendors offer a programmable middleware layer to access data from their gateway.
Figure 4. An open software platform, such as NI LabVIEW, is capable of communicating with wired and wireless sensors from multiple vendors over multiple protocols.
Click to enlarge

Many off-the-shelf wireless measurement devices only offer a means of viewing data within a fixed-function data logger. For very simple applications, this is sufficient. However, a turn-key approach complicates measurement system design if a vendor does not provide all the functionality required for a particular application – especially when that one vendor cannot support all the necessary sensors with their hardware solution. Consider the setup shown in Figure 4.

Company A may offer a good wireless strain sensor with an RS232 gateway, and Company B may have an efficient solution for wireless temperature measurements with an Ethernet gateway, but retrieving data from both sets of devices within one application may be impossible with off-the-shelf data-logging software. The picture becomes even more complicated if measurements over a cabled bus, such as USB, must be incorporated as well.

In light of the number of wireless protocols available, using an open software platform capable of communicating with all of them is a prudent choice. Many vendors offer a programmable middleware layer, such as a driver API or standard communication protocol, to access data from their gateway. For example, the NI-DAQmx driver software (http://zone.ni.com/devzone/cda/tut/p/id/5434) uses the same API for USB, PCI, PXI, Ethernet, and Wi-Fi data acquisition. In addition, National Instruments has partnered with several wireless sensor vendors to provide wireless sensor LabVIEW (http://zone.ni.com/devzone/cda/tut/p/id/5435) libraries through its instrument driver network. The ability to incorporate both wired and wireless measurements from multiple vendors within a single measurement system can provide flexibility for your application as well as protect your existing hardware investment.

A Glimpse into the Future

As wireless technology continues to evolve, it is only a matter of time before it becomes commonplace in measurement applications. Today, there is no one silver bullet for every wireless system – different technologies are better suited for different applications. Moving forward, however, the wireless sensor industry will likely see a consolidation of standards and an increased focus on security. In addition, less focus will be put on wireless technology and more emphasis will be placed on the quality of the measurement itself. Designing a system today in anticipation of these changes will put you in a position to take advantage of new wireless technology for years to come.

For more information, contact Charlie Stiernberg, Product Manager, Remote Data Acquisition National Instruments at charlie.stiernberg@ni.com

Resources

Topics

Advertisement

Share this Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading