Grant Stagg

I build algorithms for safe autonomous flight in contested environments. My research focuses on probabilistic engagement zones, uncertainty propagation, and real-time path planning under adversarial conditions.

1. Decentralized Sparse Gaussian Process Regression with Event-Triggered Adaptive Inducing Points

Published in the Journal of Intelligent and Robotic Systems, Aug. 2023 📄 Paper

  • Developed a decentralized sparse Gaussian process regression formulation using event-triggered adaptive inducing points.

2. Multi-Agent Path Planning for Level Set Estimation Using B-Splines and Differential Flatness

Published in IEEE Robotics and Automation Letters, May 2024 📄 Paper

  • Developed a multi-agent path planning algorithm that uses differential flatness to generate kinematically feasible and informative trajectories for level set estimation (LSE).
  • Formulated a novel objective function for LSE path optimization that balances exploration and exploitation and is differentiable.
  • Implemented a decentralized LSE path planner using a block coordinate ascent optimization strategy.

3. Probabilistic Weapon Engagement Zones

Presented at the 2025 American Control Conference 📄 Paper

  • Developed a method to prevent engagement in differential games under uncertainty by linearizing BEZ equations to form linearized PEZs.
  • Conducted a comparative analysis of linearized and Monte Carlo PEZ methods.
  • Implemented a PEZ-based path planning algorithm that ensures a bounded probability of agent neutralization.

4. Bi-Level Route Optimization and Path Planning with Hazard Exploration

Accepted for presentation at the 2025 Conference on Decision and Control (CDC) 📄 Paper

  • Developed a path planning algorithm that integrates Bayesian hazard probability with B-spline trajectory generation under kinematic constraints.

5. Cooperative Multi-Agent Path Planning for Heterogeneous UAVs in Contested Environments

Submitted to the AIAA Journal of Guidance, Control, and Dynamics (JGCD) 📄 Paper

  • Developed a cooperative path planning algorithm for low-priority agents that jointly optimizes regional exploration and uncertainty reduction to learn radar hazard distributions.
  • Designed a risk-aware path planner using weighted Voronoi diagrams for environments with fully known radar parameters.
  • Introduced a risk-aware planner using generalized Voronoi diagrams for scenarios with uncertain radar parameters.

6. Turn-Rate Limited Probabilistic Weapon Engagement Zones

Accepted for presentation the 2026 AIAA SciTech Forum; journal version to be submitted to the AIAA Journal of Aerospace Information Systems (JAIS)

  • Analytic solution for the turn-rate limited turn-straight EZ (CSBEZ).
  • Uncertainty analysis using linear, quadratic, Monte Carlo, piecewise linear, numerical integration, and machine learning approaches to create the turn-rate limited Probabilistic engagement zones (CSPEZ).
  • A rigorous comparison of these uncertainty propagation approaches for CSPEZs.
  • Path planning algorithm using a probabilistic CSPEZ safety constraint.