Less wire, more speed

John Hother, MD of consulting engineers Proneta, outlines the increasingly apparent advantages of using wireless technology in real-time data-gathering


Mobile phone networks (GSM) carry digital data all the time. As a result, they offer a very cost-effective way of carrying data from remote sensor sites to a central monitoring facility, and vice versa. Here we explore how it is done, and the benefits and drawbacks.
Our aim is real-time data-gathering, but real-time is a very ambiguous term. Data can be collected by a technician with a van, driving from site to site, and in many applications the delay in getting the data to where it is needed is quite acceptable. However, there is growing dissatisfaction with this method on two grounds - the high cost and the safety implications.

An easy way to judge the real-time requirement is to consider how quickly an alarm condition in any sensor needs to be recognised, and appropriate action triggered. Separate from such technical considerations, a dominant requirement is to keep both OPEX and CAPEX low.

One step up from the 'man in the van' is a traditional wired architecture. This is perfectly practical where the sensor sites are fixed and close together. However, as Dawn Fuller commented (WWT, July 2005), "cabling and installation costs can run as high as 80% of total system cost, and once a cable is installed it's costly and time-consuming to relocate, even if it needs to move only a few feet".

In a traditional wireless scheme, antenna masts provide radio signal coverage for the areas where the sensor sites are located. Each mast needs antennae for local and trunk paths, plus transmitter/ receiver amplifier and power supply equipment, and of course a local supply of electricity. The complexity of the masts, and the need for land on which to site them, leads to significant costs. If no mains electricity is available, it is possible to use solar-charged batteries. Sometimes it is not justifiable to provide masts for isolated peripheral sensor sites, in which case another sensor site in range can be equipped (with some extra complication) to provide a repeater function to bring that outlying site into the fold.

Skills to pay the bills
The planning of such a wireless system goes far beyond just selecting the right equipment and securing appropriate transmission licences - for reliable operation, a good deal of skill is required in handling the radio propagation issues at an early stage, before decisions on mast location are finalised. Masts are not entirely trouble-free, and need routine maintenance, which can be a significant extra OPEX component.
An attractive alternative is architecture using GSM/GPRS. The masts become the responsibility of the mobile phone operator, and the costs are shared with mobile phone users. Operation of the radios has changed because the mobile phone systems use a cellular architecture. Because the user does not carry the CAPEX or OPEX of the radio system directly, there is potential for large savings to be made.
Each sensor site is equipped with a GSM wireless module which works with the mobile phone network and is fitted with a SIM card. A key to making this scheme useful is the link between the mobile phone system and the internet, provided as part of the
service by the mobile phone network operator - although it is not really visible to voice users.

The monitoring station becomes a computer connected to the internet. Later we will look briefly at the special software needed within it. We will assume that the mobile phone network provides the GPRS facility, which is normal in the UK and simplifies the
operation for our purposes.

The GSM wireless module in each sensor site is programmed to send periodically a package of data to the database in the computer, containing the latest sensor values, and using that computer's internet (IP) address to locate it. The sensor sites can all do that at the same time (provided there is GPRS facility and sufficient cell capacity, which is usually the case). At any time, real-time current values at a particular site can be interrogated by the computer.

In the same way a site can also send an alarm data package at any time an alarm condition is detected, which could include site identity, site position, sensor data values, sensor data trends or current sensor limits, to make a few examples.
The computer can relay that alarm immediately to the mobile phone or PDA (as SMS text or email), of the person best placed to respond. It is also feasible for the sensor site to call directly a supervisor's mobile phone or PDA, using its own phone book. Such capabilities may not be needed initially, and could be an unnecessary
complication, but they do provide many ways to develop the facilities of the system in subsequent enhancement programmes.

Help off the shelf
From the computer end, all the sensor sites appear as automatic senders of data packages, using conventional IP addressing. For that reason, much of the software in common use is entirely compatible, and provides a highly cost-effective basis. The designated central computer needs a database to hold all the sensor values as they arrive, in an orderly and retrievable form. General-purpose software such as MS Excel can be used as the basis for this, or packages are available from specialist software houses such as OSI or Matrikon. The database needs an interface package that links it to the internet and interprets the specific data types from the sensor types installed at the remote sites. With all that installed and running, the database, as a matter of course, always contains the current and historical data values of all the sensors in the system.

Usually the best way to view and use the data is not to look directly into the database, but to have a software package that constructs web pages on demand with all the sensor data contained in easy-to-view form. Those web pages are then accessible from any connected computer with a web browser. Access may be restricted to computers directly connected to the company network, or can be routed out through a firewall to the internet, allowing any user with the necessary security passwords to access the sensor data in a user-friendly form. Now that PDAs with internet capability are readily available, this is becoming a powerful option.

All that can easily be outsourced as a turn-key service. All the user needs to access current and historical data is a computer with an internet connection and a web browser (eg IE or Firefox). The user pays a monthly fee to the provider, who in return sets up and maintains the secure web server that communicates constantly with all the user's sensor sites, and delivers all the facilities the user needs, via a password-secured website on the internet.

Alternatively, the user can provide the server and the software. The software could be in-house-designed or a bought-in proprietary package such as those from OSI (ProcessBook / ActiveView) or Matrikon. This option is likely to be more attractive to users having substantial technical IT departments. The key to retaining the option is the user insisting on open standards being used from the outset in the design of the sensor data-gathering system.

With any use of the internet and wireless communications, security is a concern. That means protecting against unauthorised access, which can be simply reading data or can be changing settings. It also means protecting against denial of service, in which a hacker deliberately disables part of a communication system by diverting a path or by overloading some point in the system.

To protect against such threats, the internet-enabled GSM system described here has several security features. Dynamic IP address assignment makes it very difficult for hackers or their software to find any address to attack, because the addresses are constantly changing. Encryption of data (in both directions) is essential whenever using a public network such as the internet. 128-bit encryption, as used in online banking, is easily incorporated because web browsers in common use support that automatically. Multi-level password access is familiar to all computer-users - different users have different passwords, and thereby access to different aspects of the system. An integral hardware firewall in each GSM wireless module provides a high degree of protection, even from a determined hacker who does succeed in finding its address. Beyond those security measures there are options to use Virtual Private Networks (VPN) and certificates, which make unauthorised access even more unlikely.

Technology for all
As an indication of the affordability of this technology, we are currently installing these systems in two high-street applications. In washing-machines in laundrettes, the remote monitoring system protects the revenue stream, reduces operating costs and protects the business's reputation for reliability. In chiller cabinets in retail outlets, the remote monitoring system provides fast, documented response to early warning of failure, protecting the value of stock goods, and protecting the business's reputation for safe food.

Here is a simple example of the very low running costs:
  • GSM connection for ten sensor sites £60/pa,
  • data charges 600MB/pa/sensor site = £300/pa/site (ten variables of 20 bytes each per site at 10sec update rate).


  • These figures are very much lower than the maintenance costs for traditional wired or wireless infrastructure (OPEX).

    The mobile phone market is dynamic, with systems rapidly becoming obsolete and being replaced. The technology is targeted at the most important customer segment - the users who provide the bulk of the revenue. So the tastes of young people dictate the development of the phones themselves. Most users are not aware of the continual changes happening behind the scenes in the networks to support those new phone facilities, for example the increasing bandwidth becoming available so pictures, video clips and even TV programmes can be sent.

    For anyone planning to use the network for data-gathering, this continuous evolution raises the question of how long a user's system will continue to be compatible with the network. The good news is that the network operators try hard to retain 'backwards compatibility', so that yesterday's phones still work with tomorrow's networks. As far as we can foresee, this will continue indefinitely, but forecasting the path of technology is notoriously difficult, so it is best to be pragmatic.

    An example of a step-change in technology that prevents backwards-compatibility was the wholesale change from analogue mobile phones to digital ones, about 15 years ago. Within a short period, analogue phones became useless. The saving grace was that the phones themselves were improving at such a rate that users chose to buy new, better ones quite frequently - and an important aspect of that turnover of equipment was the affordability of the new phones, resulting from the huge production volumes.

    The GSM units for data-gathering will probably always be much more expensive than mobile phones, simply because the production volumes are so low. However, they will also probably still remain much less expensive than alternative solutions - other wireless solutions and even cables (when their hidden costs are included).

    A sensible strategy to minimise the possible impact of changes in the networks would be to ensure that the wireless module (ie the GSM module) is a separate unit from the sensors, and can be swapped for an upgraded unit should the need arise. This also helps in the event of a failed GSM wireless module - repair can be by replacement. It is worth bearing in mind that any 'data-gatherer' user will probably invest a good deal in the IT/software aspects of a complete system, so any measures which minimise the need to modify that part when circumstances change will reap rewards in the long term.

    There are many examples of the use of internet-enabled GSM networks for data-gathering, but they are often invisible to the general public. It is indeed a pervasive technology, because it has proved to be very cost-effective, particularly where there is a need to monitor widely dispersed sites and especially if sites are relocated from time to time. It avoids any direct involvement (or responsibility) for the user in long cables or radio masts and repeaters, so installation and maintenance costs are low and flexibility is high. Implementation is quick and easy, as is relocating sites or adding new ones.

    GSM and GPRS are well-proven technologies in daily use, delivering high reliability. When linked with the internet as described, any internet-enabled device (PC/PDA/phone) can deliver all the remote monitoring facilities one is ever likely to need, and the built-in security aspects restrict access to properly authorised users.

    Acknowledgement
    This article is based on a presentation at Data Communications for the Water & Waste Treatment Industry, Bretby, 30 June 2005.


    Tags



    Topics


    Click a keyword to see more stories on that topic, view related news, or find more related items.

    Comments

    You need to be logged in to make a comment. Don't have an account? Set one up right now in seconds!


    © Faversham House Group Ltd 2005. edie news articles may be copied or forwarded for individual use only. No other reproduction or distribution is permitted without prior written consent.