With the development of new, advanced wireless networking technologies, wireless communication is becoming more attractive to industrial control applications, providing many benefits such as faster installation, physical relocation without reconfiguration, and savings from cable and installation costs in addition to eliminating the spaghetti of wires that clutter most factory environments. However, industrial controls require a robust interface solution to ensure the quality of service and security crucial in the industrial settings. Issues of security, safety, and interference still need to be resolved. This article will discuss the concerns of implementing the wireless networking solution in industrial controls vs. wired solutions.

Wireless networks have been enjoying tremendous success in recent years in corporate, residential, and hot-spot applications. Vendors such as Linksys (recently acquired by Cisco) and D-link achieved triple-digit growths even during the worst corporate spending downturn in decades. Cheaper prices are making this technology more and more attractive to many customers, even in industrial applications, where requirements are extremely rigid and rigorous.

Faster installation, physical relocation without reconfiguration, and freedom from cables make this technology attractive for new deployments and unfixed nodes. For new deployments, where enterprises often pay thousands of dollars in electrical contracting to lay out the new cable infrastructure, wireless networks provide substantial price savings by eliminating the installation cost. Although the up-front costs of a wireless LAN are higher than those of a wired LAN, once a node is moved, it is paid for. Further, wireless networks eliminate the spaghetti of wires that clutter most factory environments, eliminating risk of cable-related problems. Catastrophic system failures due to system power accidentally shorting to the communication bus or similar incidents are completely eliminated. However, even with these benefits, wireless technology may lack the basics to become the engineer's preferred choice for industrial controls. Reliability, security, and maturity of this technology, compounded with TCO (total cost of ownership) forces most engineers to think twice about the feasibility of this new technology.

In order to objectively evaluate the feasibility of wireless networks in industrial applications, the history of telecommunications and industrial controls needs to be reviewed, since the migration paths follow similar steps but with different timelines. Modern telecommunications started with the invention of the telegraph that followed Alexander Graham Bell's telephone. Around the 1960s, digital communication between trunks started replacing analog links, but humankind was still bound by wires.

In 1946, the first commercial mobile telephone system was introduced in the St. Louis area, but it wasn't until 1983 that wireless cellular phone service really started. After years of development and licensing resolution, Advanced Mobile Phone Service was launched in Chicago by AT & T on October 13th 1983. Due to bandwidth and other limitations, migration to digital systems was unavoidable. Second (2G) generation digital networks were introduced in 1992.

By the mid-1990s, the Internet boom changed history, and growth for data traffic surpassed voice for the first time. Today, most of the traffic on service provider networks is data. Although the migration to 3G and convergence between voice and data networks is delayed, various wireless networking technologies have been gaining wide acceptance. For personal area networks, Bluetooth is paving the way for non-wired short distance connections between various devices. For local area networks (LAN), various IEEE 802.11 protocols enable communication in corporate, residential and hot-spot environments.

Not surprisingly, the history of industrial networks is much shorter. By the 1940s, process instrumentation used 3-15 psi pressure signals for monitoring control devices. Around the 1960s, the first standardized communication method was introduced: 4-20mA, pure analog current loop signaling. In the 1970s, industrial applications began using PLCs (programmable logic controllers) and digital computers.

In the middle of the 1980s, the industry quest for a standardized all-digital fieldbus became a reality. However, major industrial companies and countries, mainly Germany, France, and the U.S., did not want let go of their de facto standards, so multiple competing standards came into being such as Profibus, InterBus, DeviceNet and others. These fieldbuses are simply all-digital, serial, two-way communication systems that serve as a LAN for factory/plant instrumentation, monitor and control devices. Evolution from wired analog to wireless digital communication is already happening in industrial control, but compared to digital fieldbuses, market share for wireless is almost negligible.

Communication Methods in Industrial Control

• 1940s: 3-15 psi pressure signals


• 1960s: 4-20 mA signaling


• 1970s: PLCs & Computers for monitor and control


• 1980s: Various All-Digital Field-buses


• 2000+s (???): Wireless

Reliability and maturity are the main hurdles facing wireless networks in industrial applications. Current wireless technologies are based on the TCP/IP (Transmission Control Protocol/Internet Protocol) that was originally designed to transport packets using the best-effort delivery paradigm in a connectionless and unreliable communications network infrastructure. The TCP enables reliable connection-oriented transport over the unreliable IP network. Unfortunately, the IP network drops packets due to transmission errors or during congestion without notifying the upper layer protocols. This unique nature of TCP/IP provides adaptability to any network condition, but it creates wide variations in packet delivery times known as network latency. This delay is unacceptable to most industrial applications, so TCP/IP alone cannot be used for synchronous data collection from sensors and delivery to actuators at set intervals, mostly in the tens of milliseconds range.

In addition TCP/IP is not designed for five-9 uptimes (99.999%), which is a hard requirement for many mission-critical industrial applications. Internet Engineering Task Force, the standards body responsible for TCP/IP, has been improving TCP/IP, and they have been working on various signaling mechanisms to provide traffic engineering on a per flow or connection basis. One example of such work is RSVP (Resource Reservation Protocol), which enables applications to signal flows that require very low latency and specific bandwidth requirements. Routers and switches use this protocol to reserve resources and provide transport for time sensitive traffic.

Another protocol is DiffServ (Differentiated Services) that uses the header portion of an IP packet to stamp classification information related to a service level agreement. These stamps provide routers a guideline to reserve required bandwidth and latency requirements. Unfortunately, none of these protocols really addresses the requirements for industrial applications. Current wireless solutions from a number of vendors provide workaround solutions by wrapping standard protocols around proprietary ones, enslaving companies to a single vendor with no price or feature competition.

Security is another major problem that needs to be resolved before widespread use of wireless networks in industrial applications. Because all wireless networks use air as the medium to broadcast their data, they are vulnerable to eavesdropping, where a sensitive radio and a directional antenna can listen in on company traffic from a nearby location. Companies must implement proper encryption mechanisms to protect sensitive data, especially their intellectual property, on which they spend billions of dollars to gain competitive advantage. Keys used to encrypt the data need to be updated on a regular basis, adding significant challenges in terms of man-hours and possible downtime.

Interference to and from other wireless devices is another piece of this puzzle. Because current wireless technologies use ISM (Industrial, Scientific, Medical) bands, systems need to coexist with other wireless networks. Due to the lower utilization compared to 915 MHz and cost savings over 5.8 MHz, systems are generally expected to employ 2.4 GHz range where Bluetooth and various flavors of IEEE 802.11 operate.

Sadly enough, these two systems interfere with each other when clients do not use access points to communicate. One way to prevent interference among these different networks is to use proprietary protocols similar to IEEE's protective mode, where all clients check with the access point before transmitting data, to ensure the channel is clear in heterogeneous IEEE 802.11 networks. For WLAN networks, this slows down the traffic, but this is simply unacceptable for industrial networks. Furthermore, other sources around the factory floor such as motors, welding machines, and industrial equipment also add to interference. Wireless signals need to be more resilient in the face of this type of interference, and extra error checking and correction may be necessary for reliable connections, which further complicates the design and increases the cost. Moreover, non-uniform coverage must be mapped and appropriate measures need to be taken either by adjusting output power or adding extra additional access points.

Spectrum Allocated for Unlicensed Industrial,
Scientific, Medical Use: ISM Bands

915 MHz

• Bandwidth: 26 MHz


• Availability: US/Canada


• Cost: Low


• Utilization: High

2.4 GHz

• Bandwidth: 83.5 MHz


• Availability: Worldwide


• Cost: Medium


• Utilization: Medium

5.8 GHz

• Bandwidth:125 MHz


• Availability: US/Canada


• Cost: High


• Utilization: Low

Finally, the financial side of wireless networking needs to be put under a microscope by carefully examining the TCO. Return on investment calculations used to justify the benefits of wireless networks need to include training for the personnel, software updates for patches, key updates, and downtime. The learning curve involved with wireless networks is fairly steep, quite different from the wired world.

This article only touches the surface of issues that are new and not well understood. In addition to parameters common to the wired world such as throughput, latency, and error performance, completely new parameters unique to wireless networks must be considered. Radio-signal impairments such as fading, interference, Doppler effects, reflections, and multiple paths need to be evaluated, in addition to power management, MAC (Media Access Control) layer collision and avoidance. Wireless networking in industrial control is well suited for non-mission-critical nodes, but crucial data transmission between sensors, actuator and control/monitor nodes should be evaluated cautiously.