Feature change in osi3::EnvironmentalConditions

[AVLRoding]

1. Modification on attribute: Precipitation

• Description: The actual precipitation intensity attribute consists in an enumeration (PRECIPITATION_UNKNOWN = 0 , PRECIPITATION_OTHER = 1 , PRECIPITATION_NONE = 2 , PRECIPITATION_VERY_LIGHT = 3 , PRECIPITATION_LIGHT = 4 , PRECIPITATION_MODERATE = 5 , PRECIPITATION_HEAVY = 6 , PRECIPITATION_VERY_HEAVY = 7 , PRECIPITATION_EXTREME = 8), so it is a qualitative approach rather than a quantitative approach. The proposal would be to change this to a quantitative value (float or double) in mm/h.
• But why is this necessary? Although a qualitative approach might be sufficient for some applications, once you want to model weather and its effect on sensors in a more precise way, quantification is very useful. In the following plot, the effect of rain on a LiDAR reflectivity (Velodyne VLS 128) is presented. As you can, small changes in rain intensity [mm/h] leads to a big change in reflectivity. This kind of changes and analysis cannot be performed with the actual qualitative approach.
  

2. Modification on attribute: Fog

• Description: The actual fog visibility attribute consists also in an enumeration (FOG_UNKNOWN = 0 , FOG_OTHER = 1 , FOG_EXCELLENT_VISIBILITY = 2 , FOG_GOOD_VISIBILITY = 3 , FOG_MODERATE_VISIBILITY = 4 , FOG_POOR_VISIBILITY = 5 , FOG_MIST = 6 , FOG_LIGHT = 7 , FOG_THICK = 8 , FOG_DENSE = 9). The proposal would be also to change this to a quantitative value (float or double) in m.
• But why is this necessary? As it happens with rain, fog can influence a lot the performance of the sensors. The following graph shows the influence of fog on the LiDAR detected number of points (Velodyne VLS 128) depending on the visibility. It is almost impossible with the actual qualitative approach to tackle this problem because most of these visibilities just fall inside the category “FOG_DENSE”
  

3. Add new attribute: Drop Size Distribution (DSD)

• Description: Although rain intensity and fog visibility can give a good idea how the weather conditions are, they might be sometime not enough and a more detail information must be provided. Therefore, we proposed to include another optional environmental condition, Drop Size Distribution (DSD). The DSD consist of a 2D array containing information regarding the size of the drops and its velocity.

• But why DSD needed? Why is rain intensity and fog visibility not enough? The main problem with rain intensity and fog visibility is that they do not provide enough information about how the rain/fog looks like, how the particle looks like, its size, velocity, … properties influencing/causing physical effects, like backscattering, that again influence the sensor performance as shown before. This means that different DSD (different physically weather conditions) can lead and lead to the same rain intensity/fog visibility.