Interpolation algorithm requirements

From Intamap

This page gives the functional and non-functional requirements for the interpolation methods as a simple list of what the methods should be able to do, and what they can cope with! It is also useful to review the R INTAMAP set-up page, which is somewhat misleadingly named since it is mainly about the interpolation process as an activity (essentially this is an activity diagram for the interpolate use-case).

See also Web client requirements and Interpolation web service requirements.

Some terms:

• state: the variable (or variables?) which the user wishes to interpolate
• observations: the measurements obtained by a sensor, related to the state by:
• sensor model: the mapping from the state to the observations (also called the observation operator or forward model).

Functional requirements - things it can do

• Accept a series of observations and produce an interpolated map, with uncertainty quantified (i.e. assume a default error model for the observations, which are assumed to be directly of the state to be interpolated - an identity sensor model).
• Accept a series of observations, and specification of their errors and produce an interpolated map, with uncertainty quantified (i.e. assume an identity sensor model).
• Accept a series of observations, and specification of their errors, and the sensor (or observation) model, and produce an interpolated map, with uncertainty quantified.
  • uncertainty can be quantified in a number of ways, but in essence this really means a predictive (posterior) distribution or an approximation to this. The types of uncertainty we should support can seen in the Interpolation web service requirements

Steps within the interpolate process

The R INTAMAP set-up page gives a summary of this.

Non-functional requirements - things it should be

• reliability:
• usability:
• performance:
• portability:
• extensibility:

Back to System Requirements.