Bridging Perception and Reasoning: Dual-Pipeline Neuro-Symbolic Landing for UAVs in Cluttered Environments

2025-10-25

AI Summary

To address covariate shift and poor interpretability in vision models for autonomous UAV landing in unstructured environments—characterized by clutter, uneven terrain, and absence of prior maps—this paper proposes NeuroSymLand, an offline-online neuro-symbolic framework. It integrates large language models (LLMs) with human-in-the-loop knowledge engineering to construct verifiable symbolic rules; combines lightweight semantic segmentation with non-learning-based geometric constraints to generate scene graphs; and employs Scallop probabilistic logic programming for safety scoring and human-readable decision-making. Compared to purely learning-based approaches, NeuroSymLand achieves significant improvements across multiple datasets, high-fidelity simulations, and real-world UAV platforms: +12.7% accuracy in landing-zone identification, enhanced robustness against illumination variations and occlusions, and real-time inference efficiency (average latency <85 ms). The framework delivers a trustworthy, certifiable solution for safety-critical applications including emergency response, infrastructure inspection, and logistics delivery.

Technology Category

• Computer Vision: Visual Reasoning & Symbolic Representations
• Machine Learning: Neuro-Symbolic Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Autonomous landing in unstructured (cluttered, uneven, and map-poor) environments is a core requirement for Unmanned Aerial Vehicles (UAVs), yet purely vision-based or deep learning models often falter under covariate shift and provide limited interpretability. We propose NeuroSymLand, a neuro-symbolic framework that tightly couples two complementary pipelines: (i) an offline pipeline, where Large Language Models (LLMs) and human-in-the-loop refinement synthesize Scallop code from diverse landing scenarios, distilling generalizable and verifiable symbolic knowledge; and (ii) an online pipeline, where a compact foundation-based semantic segmentation model generates probabilistic Scallop facts that are composed into semantic scene graphs for real-time deductive reasoning. This design combines the perceptual strengths of lightweight foundation models with the interpretability and verifiability of symbolic reasoning. Node attributes (e.g., flatness, area) and edge relations (adjacency, containment, proximity) are computed with geometric routines rather than learned, avoiding the data dependence and latency of train-time graph builders. The resulting Scallop program encodes landing principles (avoid water and obstacles; prefer large, flat, accessible regions) and yields calibrated safety scores with ranked Regions of Interest (ROIs) and human-readable justifications. Extensive evaluations across datasets, diverse simulation maps, and real UAV hardware show that NeuroSymLand achieves higher accuracy, stronger robustness to covariate shift, and superior efficiency compared with state-of-the-art baselines, while advancing UAV safety and reliability in emergency response, surveillance, and delivery missions.

Problem

• Enabling UAVs to autonomously land in cluttered unstructured environments
• Overcoming limitations of pure vision models with neuro-symbolic reasoning
• Combining perception with interpretable symbolic reasoning for safety

Innovation

• LLMs generate verifiable symbolic knowledge offline
• Semantic segmentation feeds real-time symbolic reasoning online
• Geometric routines compute graph attributes without training