The applet allows realistic estimates to be put on achievable location accuracy of both DF and time difference of arrival (TDOA) passive location techniques.
The geometry of up to three ESM stations is adjustable by mouse-clicking them and draggingthem to the required locations.
ESM 1 and 2 stations are moveable along the x-axis, while the third station (ESM 3) can be moved off the axis, within the x-axis constraints of the other two stations.
Similarly, the target (black square spot) can be mouse-clicked and dragged to any position on the screen.
Scrollbars associated with the ESM stations control the mean and standard deviations of DF data. For TDOA, a single scrollbar controls the magnitude of all time errors; station errors are mutually random.
Truncated Gaussian statistics, random data is generated for each station to produce the plot data on the display.
The plotted orange square is the estimated mean position for the current random data set. Data statistics for the current data set are summarized on the lower right of the main display.
A drop-down Deployment menu selects betweenvarious station deployments. The second drop-down menu selects from various display presentation options.
Data All shows the ±1σ error sectors (with mean offsets) in addition to the modeled data. For TDOA data, the error sectors are color-coded to identify the stations contributing.
Other options allow numbered DF station pairs or TDOA station subsets to be investigated. Selecting entries such as: 1+2 or 1+3 & 2+3, from the Data menu the entry means that the display plots errors from DF stations 1 and 2 for DF deployments, but for TDOA deployments, TDOA location errors are from baselines 1+3 and 2+3.
Mouse-clicking in the display space updates the set of randomly generated data.
Moving the mouse in display space displays the mouse x,y position and also its distance from the target to allow the location noise extent to be estimated.

User Notes:
The black error circle is the notional circular error probability, within which 50% of the location error estimates should lie. Its radius is given by 0.59 multiplied by the sum of the rms values of the major and minor axes of the error ellipse calculated assuming zero mean Gaussian DF error statistics.
The program generates adjustable mean, Gaussian distributed angle or time delay, noise to model location errors. This approach is more realistic than the statistical method of DF location, which requires the angle variances to be small compared to the baseline subtended angle to constrain the positional errors within an ellipse.
It is noted from the plotted data density that the data distribution is rarely regular, but the Data Statistics display usually does give a fairly consistent location estimate from the 100-point displayed sample.
It is interesting to note that the use of extra stations or baselines does not always lead to better location estimates.