The paper presents a novel framework that describes local agent navigation generically as optimizing a cost function in a velocity space. We show that many state-of-the-art algorithms can be translated to this framework, by combining a particular cost function with a particular optimization method. As such, many types of pedestrian simulation models can be reproduced using a single general principle.
The implementation of this framework, named UMANS (Unified Microscopic Agent Navigation Simulator), is freely available online. This software enables easy experimentation with different algorithms and parameters. The output generated by UMANS can be visualized in 3-D by another open source project named ChAOS (based in Unity), which is inside the OCSR context from INRIA. 

The paper was presented at the I3D symposium conference that took place last Thursday September 17th. This edition of the I3D symposium was carried out online, due to the world circumstances, and all the material presented is available on their web.  Also, many different topics related to graphics were presented in I3D2020, all available on their YouTube channel.