Alert: Automation World now enhanced for the iPad and iPhone. Watch a quick video preview
Greenfield
Factory Automation
Bassett
Batch Processing
Hand
Process Automation
Reynolds
Packaging Automation
Campbell
On the Edge
Download this free 94-page Continuous Process Playbook loaded with industry expert advice on topics ranging from control systems, instrumentation, and industrial networks to energy management, security, and system upgrades.

Wi-Fi and the Future of Wireless Sensor Networks

Most of the discussion around wireless networks centers on WirelessHART and ISA 100. Why isn’t Wi-Fi considered?
print email

While developing the November 2012 feature article for Automation World on wireless sensor networking trends, I couldn't help but notice that Wi-Fi rarely, if ever, got mentioned. Cellular networks were referenced in a few instances, but not Wi-Fi.

Of course, office grade Wi-Fi equipment isn't up to snuff for industrial applications, but what about the communication protocol itself? After all, it’s clearly capable of handling massive amount of global communications.

It turns out that there are a number of people in industry who do think Wi-Fi is a viable network for industrial use.

One of those people is Mike Fahrion, director of product management at B&B Electronics, a manufacturer of rugged wired and wireless communications products including Ethernet servers and gateways, and cellular routers.

Fahrion contends that Wi-Fi is viable for a number of reasons, including:

• Interoperability. Compatibility with Wi-Fi and Wi-Fi-enabled devices is no longer something you have to think about. But when you deploy an 802.15.4-based equipment, such as Zigbee, WirelessHART or ISA100, you do.

• Though 802.15.4 has its roots in low cost end-nodes, and is a lighter stack than a networking model like 802.11, you need additional equipment. For example, to deploy Zigbee, you also need a ZigBee coordinator and a Zigbee router. So if you're looking to deploy a large wireless sensor network, the cost can rise exponentially.

• 802.15.4 represents about a 2 percent unit market share of wireless nodes. The remainder is comprised of cellular, Bluetooth and Wi-Fi. The fact is, the R&D dollars behind Wi-Fi are driving the technology forward much faster than 802.15.4

• Since the introduction of 802.15.4, the size, cost and power requirements of Wi-Fi (802.11) have decreased on a dramatic curve in comparison to 802.15.4.

• Fahrion also suggests that we consider Moore’s Law. “If you put your forward-thinking cap on, along with a quick refresher of Moore’s law, you start to favor an architecture that pushes intelligence out to the end node and eliminates specialized infrastructure,” he says. “Yes it raises the price and power requirements of the end node, but Moore’s law and market forces (such as Broadcom’s 802.11-related investment of nearly $2 billion in 2011 alone) seem to be joining forces to make that problem effectively a non-issue.”

My discussions with Mike on this topic will be explored more in-depth in an Automation World Deep Dive video to be released soon. So stay tuned to this blog and automationworld.com for more info on this coming soon.

In the meantime, feel free to weigh in with your thoughts on this topic in the comments section below.

David Greenfield has been covering industrial technologies, ranging from software and hardware to embedded systems, for more than 20 years. His principal areas of coverage for Automation World focus on technologies deployed for factory and process automation. Contact David at [email protected] or follow him on twitter @DJGreenfield.
3

Comments

Wi-Fi is being used on the plant floor for the past several years, in devices like laptop/tablet computers and handheld terminals that can have their batteries recharged every day. When a Wi-Fi device can run 10 years without replacing the battery, then it can be considered for wireless sensor networks for wireless transmitters. If not, too many resources will be spent replacing batteries. Wi-Fi, just like Ethernet, is just physical media. But beyond sharing access points, compatibility is something you have to think about. Without a standard application protocol, you could with 30 types of wireless sensors end up with 30 proprietary protocols requiring 30 software device drivers on your system, 30 different ways of mapping the PVs into the database, 30 different ways to configure the devices, 30 different ways to calibrate the sensors, and 30 different ways of diagnosing them when there is a problem. Replacing one device with another would require new driver to be installed, PV mapping to be done. Wi-Fi, like Ethernet, is not a complete networking solution. WirelessHART provides a common application protocol used by many manufacturers requiring only a single common driver on your system, a single common way of getting the PVs into the database, a single common way to configure the devices, a single common way to calibrate the sensors, and a single common way of diagnosing them when there is a problem. Replacing one device with another does not require new driver or mapping to be done. Problem solved. The common application protocol is a key success factor for the phenomenal acceptance of WirelessHART. Ethernet and Wi-Fi are great but one size does not fit all. USB and Bluetooth are a better fit for many applications, such as the mouse or keyboard. These technologies complement each other. Similarly, Wi-Fi and WirelessHART complement each other. Indeed, HART-IP is an application protocol for Ethernet and Wi-Fi media, today used to tie WirelessHART gateways back to the DCS. This is how these technologies complement each other.

A couple great points in the comment from Jonas. First, battery power is important in a large percentage of WSN nodes. With the newest generation of wifi modules that hit the market in 2012 targeting sensor and M2M applications, it looks likely that we have already arrived. Second, there's more to an application than just the physical and network layers. No web without http, no email without smtp, etc. That's solvable, and the sooner it's embraced by developer communities the shorter the era of confusion will be. There's not a one-size-fits-all solution lurking around the corner yet - but I'd suggest that we're getting close to one-size-fits-most. In the meantime, it feels eerily similar to the day's of staunch denial that Ethernet could ever be used in manufacturing networks.

Thanks Mike. Keep in mind the sensor consumes more power than the radio. So the D-cell must power both sensor and radio for many years. Correct. HTTP for web, SMTP for email, and HART-IP for process instrument data True, Ethernet is indeed used in automation for well over ten years (its already here), mainly at the HMI, PLC, and motor drive level, but not much at the sensor level where other sensor networks are much more common.

Add new comment

Newsletters

Don't miss intelligence crucial to your job and business!
Click on any newsletter to view a sample. Enter your email address below to sign up!
Each newsletter ranges in frequency from once per month to a few times per month at most.
The best of the essentials!
Secrets to Automation Project Success

Sign up to receive timely updates from our editors and download this FREE Automation Project Survival Guide. It’s packed with field-tested best practices from industry experts that can help make your next automation project a success.

x