Risk Assessment of Obstacle Collision for UAVs under off-nominal conditions

Banerjee, Portia; Gorospe, George · 2020 · Crossref

DOI: 10.36001/phmconf.2020.v12i1.1194

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the critical need for robust risk monitoring frameworks to enable the safe integration of unmanned aerial vehicles (UAVs) into low-altitude airspace, particularly for beyond visual line-of-sight operations. While existing safety assessments primarily focus on ground casualty risks from uncontrolled descents, this study focuses on the underexplored risk of obstacle collision caused by off-nominal conditions, such as wind gusts or degraded controllability. The authors aim to provide a predictive framework that allows for risk-informed decision-making and timely mitigation of unsafe events before a collision occurs. The proposed framework computes the risk of obstacle collision ($RISK_{obs}$) as a function of the probability of collision ($p_{col-obs}$), the vehicle’s exposed area ($A_{exp}$), and the probability that the collision location represents an obstacle ($p_{obst}$). The likelihood of trajectory deviation is modeled using a Bayesian Belief Network (BBN) that processes on-board sensor data, including GPS status, battery health, and wind speed. To quantify the probability of collision, the authors implement a kinematic 3-degree-of-freedom point-mass model. This model calculates trajectory deviations based on wind velocity vectors and the UAV’s control response time. The probability of collision is determined by simulating stochastic wind fields and identifying instances where the deviated trajectory intersects with obstacle boundaries. The framework is validated using real flight data from a DJI S1000 octocoper conducted at NASA Langley Research Center. The authors present three case studies demonstrating the framework’s sensitivity to various factors. First, they show that wind magnitude and direction significantly alter risk severity; for instance, specific wind vectors increased the normalized risk severity to 0.8, while others kept it below 0.1. Second, the study demonstrates that larger UAVs incur higher collision risks due to their greater exposed area, assuming equal control response times. Third, the authors illustrate that degraded maneuverability, represented by an increased control response time, elevates the risk of collision. The results are visualized through normalized risk severity plots along the flight trajectory. The significance of this work lies in its ability to predict collision risks proactively, enhancing the safety assurance tools required for widespread UAV integration. By incorporating vehicle health prognostics and environmental disturbances into the risk assessment, the framework supports more reliable autonomous operations. The authors note that future extensions will include penetration models to correlate vehicle kinetic energy with obstacle material properties, further refining severity computations. This approach provides a foundational method for managing risks in complex urban environments where static and dynamic obstacles are prevalent.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• telematics crash prediction