Authority-Governed UAV Platform (HMAA-UAV)

National Importance of Authority-Governed UAV Autonomy

Unmanned aerial vehicles are rapidly proliferating in defense, infrastructure inspection, disaster response, and contested-environment reconnaissance. Current UAV autonomy systems lack formal governance mechanisms that dynamically regulate flight authority based on computed sensor trust, creating vulnerabilities when operating under GPS denial, RF jamming, or adversarial sensor manipulation.

Authority-governed flight autonomy addresses this gap by making trust assessment, authority computation, and recovery enforcement first-class components of the UAV control stack. Research in resilient UAV autonomy has been identified as a strategic priority by organizations including DARPA, NASA, and the U.S. Air Force Research Laboratory. DoD Directive 3000.09 explicitly requires autonomous weapons systems to maintain appropriate levels of human judgment over the use of force.

Research Problem

Autonomous UAVs increasingly operate in environments where GPS denial, sensor spoofing, RF jamming, and adversarial interference threaten flight safety. Current commercial autopilot systems use simple threshold-based failsafes (return-to-home, emergency land) that lack formal authority governance. They cannot dynamically degrade autonomy based on computed trust or enforce graded authority constraints on flight behavior.

This project extends the authority-governed autonomy concept from ground vehicles (Project 1: Rover Testbed) to aerial platforms. The UAV implements the same SATA/HMAA/CARA governance pipeline, demonstrating that authority-governed autonomy is platform-independent and applicable across robotic domains.

Governance Architecture

The UAV implements a complete authority-governed flight pipeline. Every autonomous flight command passes through trust evaluation, authority computation, and recovery logic before reaching the flight controller.

Autonomous Drone Swarm Governance Platform

This project extends authority-governed autonomy from single-agent systems to coordinated multi-agent aerial platforms. Building directly on the rover and UAV testbeds, each drone operates as an individually governed autonomy node using SATA trust evaluation, HMAA authority control, and CARA recovery logic, while participating in a mission governed by a higher-level authority layer.

Unlike conventional swarm systems that assume continuous participation of all agents, this platform introduces trust-conditioned participation, where each drone's ability to contribute to the mission is dynamically constrained by its current trust and authority state.

Figure 1. Autonomous Drone Swarm Governance Architecture. The system implements a multi-layer authority-governed autonomy framework in which mission objectives are translated into swarm-level decisions through a mission authority node and a distributed governance layer. Each drone operates as an independently governed autonomy node using SATA-based trust evaluation, HMAA authority computation, and CARA recovery enforcement. Local trust and status information are continuously propagated upward, enabling swarm-level trust aggregation and participation control. When a drone experiences trust collapse due to sensor inconsistency or adversarial interference, its authority is reduced or revoked, triggering isolation and recovery behaviors while the remaining agents dynamically reconfigure to maintain mission continuity.

Dual-Layer Governance Model

Each drone continuously evaluates its own sensor trust and computes a local authority level. These local states are transmitted to the mission authority node, which evaluates swarm-wide conditions and determines whether individual drones may continue participating, must operate under constrained authority, or must be removed from the active mission.

Swarm Experimental Program

1. Compromised Drone Isolation: Inject GPS spoofing on one drone. Measure trust collapse, authority revocation, swarm exclusion, and role redistribution among remaining agents.
2. Swarm Degradation Under RF Loss: Remove multiple drones from communication. Measure mission continuity and objective reduction under reduced authority.
3. False Consensus Defense: One drone reports incorrect position or threat data. Evaluate cross-agent trust disagreement detection and influence suppression.
4. Recovery and Reintegration: Restore trust in a previously compromised drone. Measure controlled reintroduction under constrained participation before full authority restoration.

This platform demonstrates that authority-governed autonomy is not limited to individual systems, but can regulate coordinated autonomous teams operating under uncertainty, adversarial interference, and partial system compromise. The approach proposes a framework for swarm governance in which autonomy is continuously conditioned on trust, authority, and recovery state rather than assumed by default.

This architecture addresses a critical gap in current autonomous systems, where multi-agent coordination lacks formal mechanisms for trust-aware participation control and structured recovery under adversarial conditions.

System Schematic

Full electrical schematic showing flight controller, AI companion computer, sensor, actuator, power, and module interconnections. Color-coded by node type: blue (MCU), teal (Sensor), orange (Actuator), yellow (Power), green (Module), purple (Display).

Governance Simulation Environment

The HMAA-UAV simulation environment provides a controlled experimental platform for evaluating authority-governed flight autonomy under adversarial and degraded conditions. The simulator executes the complete SATA trust fusion, HMAA authority computation, command gating, and CARA recovery logic in real-time with configurable fault injection.

This simulation demonstrates executable validation of authority-governed autonomy rather than conceptual design alone. The simulation environment mirrors the real UAV hardware architecture, enabling direct transfer of validated governance behaviors from simulation to physical flight testing.

Experimental Simulation Environment (Research Use). This environment serves as a primary validation layer for testing authority transitions and recovery behaviors, enabling repeatable experimentation prior to real-world flight deployment.

Sensor-Anchored Trust Assessment (SATA)

The UAV trust evaluation subsystem computes a continuous trust scalar from multi-sensor fusion across GPS, LiDAR, camera, IMU, barometer, optical flow, radar altimeter, and UWB inputs using weighted Dempster-Shafer belief functions with cross-sensor validation:

Trust(s_i) = weighted belief function with cross-sensor consistency checks, disagreement penalties, asymmetric decay (fast) and recovery (slow), and single-sensor veto capability

For UAV operations, trust fusion is critical because GPS spoofing can create false position confidence while visual odometry and UWB provide ground truth. The SATA engine cross-validates localization sources and penalizes inconsistencies, ensuring that spoofed GPS cannot maintain high trust when other sensors disagree.

Authority State Machine

Hardware Architecture

The UAV uses a dual-compute architecture separating flight control (Cube Orange+ autopilot) from AI governance (Jetson Orin NX), connected via MAVLink over UART.

Key Components

Validation Metrics

Experimental Program

Five core experiments designed to validate trust-governed flight autonomy under adversarial and degraded conditions.

1. GPS Spoofing Governance: Inject false GPS signals. Measure GPS trust drop, cross-validation with visual odometry, authority A3 to A2 transition, and navigation mode switch.
2. Camera Occlusion Test: Partially block camera during waypoint mission. Verify vision confidence drop, trust fusion degradation, speed limit enforcement.
3. RF Communication Loss: Sever telemetry link. Verify RF trust collapse, authority reduction, CARA return-safe or hover-hold activation.
4. IMU/Sensor Disagreement: Inject IMU drift conflicting with GPS and optical flow. Measure cross-sensor inconsistency penalty and authority downgrade behavior.
5. Compound Failure (Critical): Simultaneous degraded vision + weak GPS + RF loss. Verify rapid authority collapse, CARA activation, controlled emergency landing.

Project Status

The platform design and specification are complete. Hardware procurement and implementation are currently underway.

Project Documentation

Complete engineering documentation for the HMAA-UAV Authority-Governed Drone Platform.

Reproducible Research Artifacts

This project provides reproducible artifacts enabling researchers to replicate the authority-governed UAV experiments and system architecture. All documentation, schematics, and specifications are available for download.

Power Architecture

The UAV platform operates from a 6S LiPo battery (22.2V nominal) with regulated power distribution for compute, sensors, and motor controllers.

Trust & Verification Chain

The UAV platform integrates SATA-HMAA-CARA governance with Cube Orange+ autopilot safety systems via MAVLink/HIL bridge.

Bill of Materials: ~$4,200

52-component UAV platform with MAVLink/HIL integration. Estimated platform cost ~$4,200. All COTS components.

Physical Specifications

Future Work

The HMAA-UAV extends the authority-governed autonomy research program from ground to aerial platforms.

About This Project

This project is the second platform in the authority-governed autonomy research program by Burak Oktenli at Georgetown University (M.P.S. Applied Intelligence). It extends the SATA/HMAA/CARA governance architectures from ground vehicles to aerial platforms, demonstrating platform-independent authority governance. The next phase scales this architecture from single-UAV governance to multi-agent drone swarm authority coordination, where each drone's participation is conditioned on trust and recovery state.

Related: Project 1: Rover Testbed · Full Research Portfolio → · Zenodo Record → · Repository →