Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Human-like Semantic Navigation for Autonomous Driving using Knowledge Representation and Large Language Models

Augusto Luis Ballardini, Miguel Ángel Sotelo

TL;DR

The paper addresses the challenge of navigating autonomously in dynamic urban environments where fixed maps may be unreliable. It proposes a semantic navigation framework that translates informal driving instructions into declarative ASP rules, enabling non-monotonic, explainable planning with an intrinsic knowledge base of traffic rules and a dynamically generated extrinsic knowledge base. Through experiments with multiple Large Language Models, the authors demonstrate that LLMs can produce syntactically valid DLv code that yields coherent answer sets, while also revealing limitations in consistency when guidance or rule bases are incomplete. The work contributes to the field by decoupling navigation from cartography, enhancing explainability via logic-based reasoning, and outlining avenues for automatic rule extraction and richer knowledge integration in future autonomous driving systems.

Abstract

Achieving full automation in self-driving vehicles remains a challenge, especially in dynamic urban environments where navigation requires real-time adaptability. Existing systems struggle to handle navigation plans when faced with unpredictable changes in road layouts, spontaneous detours, or missing map data, due to their heavy reliance on predefined cartographic information. In this work, we explore the use of Large Language Models to generate Answer Set Programming rules by translating informal navigation instructions into structured, logic-based reasoning. ASP provides non-monotonic reasoning, allowing autonomous vehicles to adapt to evolving scenarios without relying on predefined maps. We present an experimental evaluation in which LLMs generate ASP constraints that encode real-world urban driving logic into a formal knowledge representation. By automating the translation of informal navigation instructions into logical rules, our method improves adaptability and explainability in autonomous navigation. Results show that LLM-driven ASP rule generation supports semantic-based decision-making, offering an explainable framework for dynamic navigation planning that aligns closely with how humans communicate navigational intent.

Human-like Semantic Navigation for Autonomous Driving using Knowledge Representation and Large Language Models

TL;DR

The paper addresses the challenge of navigating autonomously in dynamic urban environments where fixed maps may be unreliable. It proposes a semantic navigation framework that translates informal driving instructions into declarative ASP rules, enabling non-monotonic, explainable planning with an intrinsic knowledge base of traffic rules and a dynamically generated extrinsic knowledge base. Through experiments with multiple Large Language Models, the authors demonstrate that LLMs can produce syntactically valid DLv code that yields coherent answer sets, while also revealing limitations in consistency when guidance or rule bases are incomplete. The work contributes to the field by decoupling navigation from cartography, enhancing explainability via logic-based reasoning, and outlining avenues for automatic rule extraction and richer knowledge integration in future autonomous driving systems.

Abstract

Achieving full automation in self-driving vehicles remains a challenge, especially in dynamic urban environments where navigation requires real-time adaptability. Existing systems struggle to handle navigation plans when faced with unpredictable changes in road layouts, spontaneous detours, or missing map data, due to their heavy reliance on predefined cartographic information. In this work, we explore the use of Large Language Models to generate Answer Set Programming rules by translating informal navigation instructions into structured, logic-based reasoning. ASP provides non-monotonic reasoning, allowing autonomous vehicles to adapt to evolving scenarios without relying on predefined maps. We present an experimental evaluation in which LLMs generate ASP constraints that encode real-world urban driving logic into a formal knowledge representation. By automating the translation of informal navigation instructions into logical rules, our method improves adaptability and explainability in autonomous navigation. Results show that LLM-driven ASP rule generation supports semantic-based decision-making, offering an explainable framework for dynamic navigation planning that aligns closely with how humans communicate navigational intent.

Paper Structure

This paper contains 9 sections, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Proposed Method
  4. Semantic Navigation using ASP
  5. Knowledge Engineering with LLMs
  6. Experimental Validation
  7. Experimental Setup
  8. Results on Structured Navigation Instructions
  9. Discussion and Future Work

Figures (5)

  • Figure 1: Overview of the proposed pipeline. The figure shows two flows for generating ASP rules: a classical path via a programmer and an LLM-assisted path. This work focuses on automating this translation to overcome the knowledge engineering bottleneck.
  • Figure 2: Scene interpretation error. The vehicle control system fails to correctly recognize the intersection shown on the map and, as a result, enters the left-hand road in the wrong direction. Images adapted from FSD2021.
  • Figure 3: Possible scenarios to reach a gas station, given the instructions.
  • Figure 4: Possible scenarios derived from "turn left at the next junction, then go straight, and finally turn right".
  • Figure 5: Experiment ID3. The black car indicates the starting position. Green arrows represent the two navigation scenarios modeled by the DLV-generated rules.