There are examples of simple walking using NetLogo (see https://ccl.northwestern.edu/netlogo/index.shtml) in the model library that comes with the software. More elaborate models can be found here: https://www.comses.net/codebases/?query=walk
Both of links were very good. Especially second link.
The samples in second link make a basic understanding about agent based modeling, but the link do not describe with details. On the other hand, I would like to find some sources describing agent based modeling with details. Especially about prediction walking behavior.
URDME is very good, have a look. http://urdme.github.io/urdme/
It is a Monte Carlo sampling algorithm solving Reaction-Diffusion Master Equations, treating diffusion as events where an agent cross the border of some two neighboring subvolumes.
That was very useful. I found some important point in it. However, I think URDME is not appropriate to spatial simulation. Actually, I will do spatial agent based modeling. In addition, it is important for my research that the software links with GIS.
We have implemented the Social Force model in AnyLogic; a presentation on the topic is attached. Let me know if you are interested in more material or sample models.