search

MCZhi / Driving-IRL-NGSIM

[T-ITS] Driving Behavior Modeling using Naturalistic Human Driving Data with Inverse Reinforcement Learning
MIT License
191 stars 37 forks source link

about the regulation.py #10

Open siami233 opened 1 month ago

siami233 commented 1 month ago

dear author: I mentioned in your code,a file named "regulation.py" is not used elsewhere， in your paper, you do not mentioned ahout it,so what is the usage of this file?

siami233 commented 1 month ago

Another issue is that, while your proposed personalized reward function for each driver is more accurate than a general reward function established from human driving data, the general reward function can be directly used by autonomous vehicles for decision-making and planning once established from human driving data. However, for the personalized reward function you proposed, each human driver has a different reward function. In simulation or real-world driving, which human driver's reward function should the autonomous vehicle choose as a reference for decision-making and planning?