Crowd groups with vision and social forces

[douglasbandeiraivo]

Hi,

I'd like to do an implementation of crowd simulation using Menge framework and I'm
not very sure about how to do it. Initially, I was willing to implement something
similar to the work of Jaklin et al., 2015 (Adding Sociality to Virtual Pedestrian
Groups). In this work there are some key points:

1 - when the agents of the group are close enough (when there is coherence),
all members of the same group share the same global path. Should I extend NavMeshVelComponent and extend setPrefVelocity() for this?

2 - From the preferred velocity of the agent, a new preferred velocity
is generated through a collision avoidance model based on vision. Can I do it extending a VelModifier ou shoud I do it extending BaseAgent and implementing computeNewVelocity()?