Wireless has been around for a long time, but not in the form we’re familiar with today. Elements, societies and animals all communicate by using a combination of chemical, odour, sight, acoustic and electromagnetic means. The ability to talk to other “directly connected” groups is what drives life and evolution. What do I mean by directly connected? Our brain calculates and communicates with the rest of our body by biological wires. Each of us then communicates by voice (wireless). We have extended our range with radio and telephone, but still use the I/O of our body.
This is a perspective from my farm in New Hampshire, where we’ve installed our own wireless network.
“The Barn” is located in a small town (1,200 people) in the woods, 50 miles from Boston. We are at the end of the power line and don’t have cable or DSL, although we do have satellite TV. For years I used dial-up, but hungered for a one-megabyte T1 line. I spent tons of money and installed two meshed access points to get my bandwidth. Now, I am wireless.
We run a T1 to an access point for broadcast access over a wide region. We also feed another access point across the valley, which then beams it to me. The signal goes from the roof antenna to our room router. Local access is distributed across The Barn via 802.xxx. Each of these jumps requires different frequencies and bandwidths. There is no single specification that can work for all.
In process and discrete manufacturing, we need similar communications. Originally, this was carried out by people and paper. Now we use copper and radios. The biotech equivalent of our local brain is the mixing tank, CNC controller or programmable logic controller. These local brains have direct control over the factory unit and need to communicate status and receive data from other units. Each segment needs net communication and a closed loop for local optimization – just like my barn.
There is no doubt that wireless technologies have improved communications in the manufacturing industry, but let’s take a journey into the future. A long-term view of wireless technologies is necessary for establishing standards for our industries. These standards need to take the clustered needs of a system and allow each segment to easily talk to one another. The specs need to be able to describe the links between wireless groups and have less to do with each element on the network. Claude Shannon’s 1948 paper, A mathematical theory of communication, established the basic theory used today. But more recent studies suggest that we need to use more of this science in our march to tomorrow. The fundamental technology is to have multiple independent systems communicate easily.
Clustering is an example of one approach to wireless. Humans have high-speed internal management systems, and very low wireless (speech) communication. Our internal system communicates by hormones, electricity and chemicals. Each of these internal sets has different rates. Externally, we use speech and touch. The lower external performance nicely connects humans together. We do not connect my liver to your liver. This grouping inside our bodies localizes organ performance, while the longer range speech connects the bag of organs to another human.
In a similar manner, telephones, cell phones, personal computers, television, sports and gaming are all examples of local optimization with a limited cross group contact. This “object-oriented” approach to communication allows users and innovators to make disruptive internal changes without isolation from the net.
For instance, my house has an oil heating system. The furnace has local loops, communicates with the low bandwidth thermostat in the kitchen and responds to the thermostat condition. Imagine that the igniter in the furnace is directly connected to a central system as well as the thermostat. All sensing would be passed onto the Internet and a master computer would make all of the decisions. A top-down approach would connect the house thermostat to an ISP, and the burner to the ISP. Command and control via the Internet does not follow the simplistic biological clustering mode. This is a faulty, top-down approach to wireless that seems to be an engineering trend.
Clustering allows 10-cm wireless systems with gigabit rates. Printed circuit boards will no longer exist in physical form. The topological connectivity will be virtualized with the 10-cm wireless capability. Each semiconductor will have a nanotooth capability to communicate across a basket of processors, powered by wireless power transmission. We expect the future of wireless will be in the direction of extremely small systems and clustering.
On the topic of standards, many organizations are working on wireless standards, but not all of them, in my opinion, will benefit the industry. I believe the new specification for industrial communication protocol, the Common Industrial Protocol (CIP), from ODVA, is an example of faulty thinking.
The CIP is designed to provide users with a unified communication architecture throughout the manufacturing enterprise, and to allow users to benefit from the advantages of open networks. The protocol includes coherent integration of I/O control, device configuration and data collection. I think it should include coherent segmentation instead. The CIP also includes the seamless flow of information across multiple networks. I think it should be managed flow. It’s comprised of multi-layer networks without the added cost and complexity of bridges and proxies, but I think it should use bridges to increase flow. The protocol also includes the freedom to choose best-of-breed products with the assurance of competitive prices and low integration cost. I believe in the freedom to choose, but I think this approach hinders it. It also includes a single, media independent platform for all CIP networking technologies, while I think there should be multiple platforms with bridges. The one thing we do agree on is a minimized investment in system engineering, installation and commissioning.
The clustering and “small distance, high bandwidth” focus described here is not the only direction. Other groups are focusing on adaptive routing, frequency hopping, mesh, spread spectrum, endpoint redundancy, artificial life, cognitive nets and autonomous compression.
Clustering also allows innovation on a local basis while still servicing the larger networks. Engineers need to consider transmitting only the needed information. Think small and use the biological model.
Dick Morley is the inventor of the PLC, an author, speaker, automation industry maverick and a self-proclaimed ubergeek. E-mail him at morley@barn.org.
