Product Releases

Sensing the Advantages of a Mesh Network

Mon, 01/05/2009 - 5:10am

One of the most interesting trends in the wireless sensor industry is the emerging development and synergy of energy harvesting and ultra low power wireless mesh networking. By using a low power wireless mess network combined with energy harvesting, it is now possible to develop inexpensive and extremely reliable ZigBee-based sense and control networks that can cover vast areas.

Cees Links, CEO & Founder, GreenPeak
Many are familiar with the so-called star networking concept based on the architecture of hubs, routers, gateways, etc, that dominates most home, business, and industrial networks. However, there is another network architecture that is becoming popular—mesh networking.

Mesh Networking

In a mesh network, each node can talk with every other node in the network that is within range and without the requirement of going through a central base-station. One of the nodes can be dedicated to maintain connection to the "outside world" and serve as a gateway, but even that node is just "one amongst equals"—only it is dedicated to performing a special extra function.

Wireless mesh networks can be comprised of energy harvesting powered transceivers that enable sensors to transmit info to the coordinator even when a direct path is not available.
One of the more appealing features of a mesh networking configuration is its self-forming, self-healing character; with the help of a smart algorithm (and/or dumb tables), information finds its way ("hops" or "is routed") from source to destination along a path of intermediate nodes. If that path is "blocked" (e.g., an intermediate node has dropped out or the infrastructure has changed), that same algorithm finds an alternative route (assuming it exists). Although this type of architecture may sound very beneficial and logical for business, home, and industry networks, in fact, mesh networking topologies are actually quite rare.

Interestingly, the Internet that most rely on for daily communications and business, functions as a giant, international mesh network, as data from one user travels across the world, following different nodes and gateways as needed to create a path that delivers the bits as quickly and reliably as possible. Because of the backbone's mesh network, the Internet is very resilient and reliable.

However, because of power issues, creating a wireless mesh network is a bit more complicated. The nodes in a wireless network are usually running of batteries and that battery life is limited. For a mesh network to work effectively, all the nodes must be "on" all the time, which usually drains the battery so quickly that meshing becomes impractical. Furthermore, hopping over multiple nodes can also introduce "latency" and occupation of precious bandwidth. For example, it is clear that when making a telephone call, one would prefer to be directly connected to a base station versus hopping via a set of intermediate cell phones.

Conversely, for most sense and control networks, latency is not a primary concern and the amount of data transmitted is usually so low that hopping does not waste precious bandwidth. Therefore, it is not surprising that ZigBee has included mesh networking in its standard. However, the current ZigBee specifications only define the routing nodes as powered routers. Because energy consumption is usually an important issue for ZigBee mesh network nodes, the ZigBee backbone has to be powered and the associated sensor nodes implemented in the traditional star networking pattern.


There are a wide range of current and emerging sensor applications that truly require a mesh network. These include large area applications such as agriculture, foresting, or larger building structures. Because of the expense of power lines and data communication lines, it is impractical to implement a base station infrastructure for these applications that can cover multiple square kilometers.

A network of environmental sensors connected by an ultra low power wireless network can be used to help prevent forest fires.
Instead, the solution is a combination of three technologies: ultra low power radios, synchronized networking, and energy harvesting. A proper system development enables energy harvesting to deliver enough energy for a node to come up (and to do its job) at regular time intervals while being asleep the rest of the time. It is important to realize that a node usually "works" for only 100 milliseconds and then can be asleep for 900 milliseconds, or even for 5 minutes.

In this way, a forest can be populated with sensing nodes that interconnect via each other (in a meshing sense) to a central node for reporting values to the outside world without the need for a powered infrastructure.

For example, it is now possible to control forest wildfires by building a network of predictive and fire sensors (using moisture, temperature, and fire alert sensors), connected by a GreenPeak low power routing (LPR) sensing network, all powered by Voltree energy harvesting. Voltree harvests metabolic energy from trees and converts it to useable electricity, in sufficient amounts to power a wireless mesh sensor network. Voltree's weather resistant "bio-energy converter" power module parasitically harvests metabolic energy from any large plant without significantly harming it with a useful lifetime of the device that is only limited by the lifetime of the host.

Ultra low power wireless networks can be powered by using Voltree's weather resistant "bio-energy converter" power module that harvests metabolic energy from any large plant.
With the proper design and dimensioning of this type of network, forest wildfires can be quickly identified over a very large wilderness area. Instead of having to wait for a fire to grow big enough to be noticed by human monitors, by using energy harvesting powered remote sensors in a ZigBee based wireless mesh network, it is now possible to quickly and efficiently respond to small forest fires before they turn into major environmental disasters.


Share this Story

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