This blog briefly reviews some human driving modeling methods to help clarify the modeling thinking.

Human Driving Behavior Modeling

A Simple Roadmap

graph LR;
   A["<b>Human Driving Behavior Modeling Methods</b>"]-->B(Rule-based);
   B-->E["Car-following model (Longitudinal)"];
   B-->F["Lane-changing model (Lateral)"];
   B-->G["Gap-acceptance model (Junctions)"];
   A-->C(Cognitive psychology);
   A-->D(Data-driven);
   D-->H["Statistical learning"];
   H-->I[Clustering];
   H-->J[Classification];
   H-->K[Regression];
   D-->L["Deep learning"];
   D-->M["Reinforcement learning"];
   M-->N["Imitation learning"];
  
  style A fill:none,stroke:#333;
  style B fill:none,stroke:#333;
  style C fill:none,stroke:#333;
  style D fill:none,stroke:#333;
  style E fill:none,stroke:#333;
  style F fill:none,stroke:#333;
  style G fill:none,stroke:#333;
  style H fill:none,stroke:#333;
  style I fill:none

Lu H, Liu Y, Shen S, et al. A Comprehensive Survey on Human-like Driving: Breadths and Depths Explored[J]. Available at SSRN 5105959, 2025.

Depth:

graph TB;
   A["<b>Depth of Human-like Driving</b>"]-->B[Human-like Control]
   A-->C[Human-like Decision-making]
   A-->D[Human-like Planning]
   C-->E[Cognitive Process]
   C-->F[Interaction Strategy]
   C-->G[Whom to be alike]
   D-->H[Global Planning]
   D-->I[Local Planning]

  style A fill:#c5d9e8,stroke:#333,stroke-width:2px;
  style B fill:#e8cdd4,stroke:#333,stroke-width:2px;
  style C fill:#e8cdd4,stroke:#333,stroke-width:2px;
  style D fill:#e8cdd4,stroke:#333,stroke-width:2px;
  style E fill:#f5f5f5,stroke:#333,stroke-width:1px;
  style F fill:#f5f5f5,stroke:#333,stroke-width:1px;
  style G fill:#f5f5f5,stroke:#333,stroke-width:1px;
  style H fill:#f5f5f5,stroke:#333,stroke-width:1px;
  style I fill:#f5f5f5,stroke:#333,stroke-width:1px;

Breadth:

graph TB;
   A["<b>Breadths of Human-like Driving</b>"]-->B[Mechanism-inspired]
   A-->C[Large Model-driven]
   A-->D[Modeling-based]
   A-->E[Learning-driven]
   D-->F[Game Theory]
   D-->G[Behavioral Preference]
   D-->H[Risk Representation]
   D-->I[Others]
   E-->J[Behavior Cloning]
   E-->K[Reinforcement Learning]
   E-->L[Inverse Reinforcement Learning]

  style A fill:#c5d9e8,stroke:#333,stroke-width:2px;
  style B fill:#9db8a8,stroke:#333,stroke-width:2px;
  style C fill:#9db8a8,stroke:#333,stroke-width:2px;
  style D fill:#9db8a8,stroke:#333,stroke-width:2px;
  style E fill:#9db8a8,stroke:#333,stroke-width:2px;
  style F fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style G fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style H fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style I fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style J fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style K fill:#c8dcc8,stroke:#333,stroke-width:1px;
  style L fill:#c8dcc8,stroke:#333,stroke-width:1px;