Example of data measurement fusion from two sensors, performance comparison

[A-acuto]

In this PR, we add an example where we perform a 1 to 1 comparison between Extended Kalman filter, Unscented Kalman Filter and Particle filter tracking performances with data coming from two separate sensors, simplistically assumed to have the same specifics, in a multi-target scenario.

[A-acuto]

Many thanks @spike-dstl for the useful comments on this example. I have made all the relevant changes as suggested. A couple of comments:

• regarding the threshold resampler, now it is in terms of number of particles (also specified at the beginning as simulation parameter). The previous choice was done to minimise the problems in obtaining a single track using the PF.
• TypeError: float() argument must be a string or a number, not 'datetime.datetime' it's really odd, in my runs while compiling I never had this issue. Let's see if now the docs are failing again or not.

[A-acuto]

@nperree-dstl , @spike-dstl Thanks for the comments and feedback. I have applied the suggestions provided, including some improvements on the track initiators (all trackers are now using MultiMeasurementInitiator) and I have included a non-negligible clutter level (which makes the example, I think, more relevant). Now the tracks have the same colors between the plot and metrics (for consistency).

I have not managed to use different colors for the detections (using Plot_measurements in Plotterly) because if I pass an external keyword for the colors (e.g. marker=dict(color='x')), it is applied on both detections and clutter, making it a bit confusing.