NEUROSYMLAND: Neuro-Symbolic Landing-Site Assessment for Robust and Edge-Deployable UAV Autonomy

2026-07-02

AI Summary

This work addresses the challenges of evaluating safe landing sites for unmanned aerial vehicles (UAVs) in unstructured environments and the lack of decision transparency by proposing a neural-symbolic approach tailored for edge deployment. The method integrates lightweight visual perception with explicit symbolic reasoning, constructing a probabilistic semantic scene graph and incorporating symbolic constraints—such as terrain flatness, obstacle spacing, and spatial consistency—to perform structured safety assessments of candidate regions. As the first study to apply neural-symbolic systems to UAV landing tasks, the proposed framework achieves efficient execution on resource-constrained hardware while maintaining robustness and interpretability. Experimental results demonstrate successful evaluations in 61 out of 72 simulated scenarios, outperforming four baseline methods; hardware-in-the-loop tests further confirm its low latency, minimal resource consumption, and high deployment feasibility.

Abstract

Safe landing-site assessment in unstructured environments remains a key challenge for autonomous UAV deployment, as vision-only learning approaches often degrade under terrain variability and provide limited transparency in safety decisions. We present NEUROSYMLAND, a neuro-symbolic landing-site assessment system that integrates lightweight perception with explicit safety reasoning. The framework constructs a probabilistic semantic scene graph from onboard visual input and evaluates candidate landing regions using symbolic constraints capturing terrain flatness, obstacle clearance, and spatial consistency, enabling structured reasoning under perceptual uncertainty while maintaining edge-feasible execution. Across 72 simulated landing scenarios spanning diverse terrains, NEUROSYMLAND achieves 61 successful assessments, outperforming four competitive baselines (37-57 successes). To evaluate deployability, we further conduct 100 hardware-in-the-loop trials with randomized initial poses, profiling end-to-end latency, stage-wise execution time, and system-level metrics including CPU/GPU utilization, memory footprint, and power consumption. Results demonstrate improved robustness and interpretability with bounded edge-resource usage. Profiling shows that symbolic reasoning contributes only a small fraction of end-to-end latency, while the main computational cost arises from perception and PSSG construction. These results demonstrate the feasibility of deploying the landing-site assessment stack on edge-constrained UAV hardware, and all source code, datasets, prompts, and symbolic rule refinement examples are released in an open-source repository

Problem

• landing-site assessment
• autonomous UAV
• unstructured environments
• vision-only learning
• safety transparency

Innovation

• neuro-symbolic reasoning
• landing-site assessment
• edge-deployable autonomy
• probabilistic semantic scene graph
• UAV safety