Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Safe and Personalizable Logical Guidance for Trajectory Planning of Autonomous Driving

Yuejiao Xu, Ruolin Wang, Chengpeng Xu, Jianmin Ji

TL;DR

The paper addresses the challenge of balancing safety, efficiency and user preferences in autonomous highway trajectory planning. It introduces the Logical Guidance Layer (LGL), which uses Highway Traffic Scenario Logic (HTSL) for scenario reasoning, integrates the Responsibility Sensitive Safety (RSS) model for formal safety guarantees, and allows personalization through user-defined logical preferences to select a local target area in S-L-T-V space. Key contributions include the HTSL formalization, a three-part LGL pipeline (scenario reasoning, evaluation, and guidance-area generation), and evidence that LGL improves safety without sacrificing efficiency while enabling user customization. The approach provides a practical pathway to formally safe, user-aware planning that can integrate with existing trajectory planners and is extendable to broader traffic environments and human–machine interaction use cases.

Abstract

Autonomous vehicles necessitate a delicate balance between safety, efficiency, and user preferences in trajectory planning. Existing traditional or learning-based methods face challenges in adequately addressing all these aspects. In response, this paper proposes a novel component termed the Logical Guidance Layer (LGL), designed for seamless integration into autonomous driving trajectory planning frameworks, specifically tailored for highway scenarios. The LGL guides the trajectory planning with a local target area determined through scenario reasoning, scenario evaluation, and guidance area calculation. Integrating the Responsibility-Sensitive Safety (RSS) model, the LGL ensures formal safety guarantees while accommodating various user preferences defined by logical formulae. Experimental validation demonstrates the effectiveness of the LGL in achieving a balance between safety and efficiency, and meeting user preferences in autonomous highway driving scenarios.

Safe and Personalizable Logical Guidance for Trajectory Planning of Autonomous Driving

TL;DR

The paper addresses the challenge of balancing safety, efficiency and user preferences in autonomous highway trajectory planning. It introduces the Logical Guidance Layer (LGL), which uses Highway Traffic Scenario Logic (HTSL) for scenario reasoning, integrates the Responsibility Sensitive Safety (RSS) model for formal safety guarantees, and allows personalization through user-defined logical preferences to select a local target area in S-L-T-V space. Key contributions include the HTSL formalization, a three-part LGL pipeline (scenario reasoning, evaluation, and guidance-area generation), and evidence that LGL improves safety without sacrificing efficiency while enabling user customization. The approach provides a practical pathway to formally safe, user-aware planning that can integrate with existing trajectory planners and is extendable to broader traffic environments and human–machine interaction use cases.

Abstract

Autonomous vehicles necessitate a delicate balance between safety, efficiency, and user preferences in trajectory planning. Existing traditional or learning-based methods face challenges in adequately addressing all these aspects. In response, this paper proposes a novel component termed the Logical Guidance Layer (LGL), designed for seamless integration into autonomous driving trajectory planning frameworks, specifically tailored for highway scenarios. The LGL guides the trajectory planning with a local target area determined through scenario reasoning, scenario evaluation, and guidance area calculation. Integrating the Responsibility-Sensitive Safety (RSS) model, the LGL ensures formal safety guarantees while accommodating various user preferences defined by logical formulae. Experimental validation demonstrates the effectiveness of the LGL in achieving a balance between safety and efficiency, and meeting user preferences in autonomous highway driving scenarios.
Paper Structure (30 sections, 13 equations, 4 figures, 1 table)

This paper contains 30 sections, 13 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Logic of Traffic Scenarios
  4. Responsibility Sensitive Safety Model
  5. Preliminary
  6. Abstract Model
  7. Action Model
  8. Safety Constraints
  9. Goal
  10. Method
  11. Scenario Reasoning
  12. Initial Condition
  13. Action
  14. Solver
  15. Scenario Evaluation
  16. ...and 15 more sections

Figures (4)

  • Figure 1: The role of the proposed Logical Guidance Layer in a typical planning and control framework. The LGL performs as a middle layer between decision-making and local trajectory planning.
  • Figure 2: Our LGL (green) is inserted between decision-making (red) and trajectory-planning (purple). The input includes the goal obtained from decision-making and the current state. An extra prediction input of environmental vehicles' actions is also acceptable. The LGL finally generates the guidance area in the S-L-T-V space with the consideration of efficiency, consistency, and user preferences.
  • Figure 3: Demonstration of the operation of vehicles controlled by four algorithms in high-density environments. (a) Results of the LGL+Lattice (left to right, top to bottom). The red lines represent the planned trajectories. (b) Results of the pure Lattice algorithm. (c) Results of the RL algorithm. (d) Results of IDM+MOBIL.
  • Figure 4: The number of overrides of 3 volunteers under each user preference. The volunteers' choices are marked by red check marks.

Theorems & Definitions (12)

  • Definition 1: Road
  • Definition 2: Logical Snapshot
  • Definition 3: Logical Scenario
  • Definition 4: Tail of Logical Scenario
  • Definition 5: Syntax
  • Remark 5.1
  • Remark 5.2
  • Definition 6: Semantics
  • Definition 7: Distance
  • Definition 8: Goal
  • ...and 2 more