Simulated Network Analysis: A Guide to GeoLab's Pre-survey Analysis Tools

Planning a cost effective GPS network that meets your project goals can sometimes feel like "Black Magic". In a perfect world, your network stands alone, geometrically strong, balanced observations, equal course lengths, and consistent observations - a dream to survey. The reality is that you likely have to integrate your new control points into an existing framework (that's never where you need it), course lengths vary, your observations will be varied - in short, it's never 'a perfect world'.

You design your survey network to:

• Meet the goals of your project;
• Be cost effective;
• Check the field data; and
• Provide a measure of confidence in your survey.

As with any network, there's more than one solution, and there are some better than others. Which one to use? Each design is basically an educated guess, the results of which are never really known until the survey work has been completed, the data processed, and the observations adjusted.

What if you had a tool that would give you a sense of whether or not your design would, in theory, meet your overall project goals? Might make things easier?

GeoLab can help

By using GeoLab in simulation mode, you can create your network design and, using a 'typical' covariance matrix, perform a simulated adjustment to give you an overall sense of how well your planned network of observations will stand up in the real world. By identifying weak areas, you can plan additional observations to strengthen your network while the crews are in the field the first time rather than sending back for more. Less field time, more profit?

Simulated Covariance Matrix

For this example we've selected 3 different covariance matrices from baselines produced by one of the GPS manufactures' software and using their "sample" data sets. Using their guidelines, we reviewed the processing results and selected 3 baselines of 'average' quality produced from varying occupation times (8, 11, and 36 minutes) of dual frequency data on short distances (1100m, 1750m and 850m respectively) baselines. Each of these 'sample' matrices can be used in varying combinations to aid in your pre-analysis.

The following steps outline the overall procedure for performing a simulation:

• Prepare a network plan of your proposed GPS survey including all proposed observed baselines.
• Create a coordinate list of the proposed stations. The format can be X, Y, Z, any map projection you choose, or a standard Point name, Latitude, Longitude, Height, commas separated values, decimals of a degree. Coordinates can be scaled from a map, created from any one of your existing software packages or exported from an existing database.
• To create your own 'typical' covariance matrix for the simulation, go back to one of your previous GPS projects that employed a similar receiver used in a similar fashion to what you propose for this network. In fact, over time, you can create an entire library of 'typical' observations, single or dual frequency data, long or short baselines, varying occupation times and quality indicators, etc. etc.
• Import the coordinates of your proposed network into GeoLab.
• Use this sample IOB file as a guide to creating your own input observations file. Each course in your network must be uniquely defined (multiple observations can be included as well, just like in real life) to create the simulated network. Insert the sample covariance matrix exported from your processing software as required.
• Process the network. Note, you must specify that adjustment is a simulation, not a real adjustment. This can be done by setting the Adjustment/Simulation Option (Tools/Options) to Simulation, or by including a COMP SIM line in your IOB file.

Once the simulation has been performed, you can view the network diagram and 2D/3D station error ellipses as well as relative error ellipses to give you a feel for the strength of your network design.

Observations (GPS or otherwise) can be added or removed from the network giving you an opportunity to see their effect so you can gauge the overall effect on the network.

Concluding Remarks