May 04, 2026 PX4 Weekly Dev Update: Core System Stabilization and Scalability Acceleration

재생

PX4 Weekly Integrated Briefing

The past seven days of PX4 Autopilot development activity clearly demonstrated multifaceted efforts simultaneously pursuing stability, safety, and scalability within the flight control system. The GitHub repository saw active progress on numerous safety-critical bug fixes, alongside the addition of new flight controllers and IMU drivers, signifying the continuous expansion of PX4’s supported hardware ecosystem. Notably, improvements to the EKF2 (Extended Kalman Filter) and navigation modules indicate a strong focus on enhancing precise autonomous flight capabilities.

Of particular note is the active discussion surrounding the introduction of the Pixi build system, a development environment management tool, and ongoing documentation improvement efforts. These initiatives are foundational for enhancing the Developer Experience and ensuring the long-term maintainability of the project. Furthermore, an in-depth RFC (Request For Comments) concerning the MAVLink implementation reflects the community’s commitment to exploring future directions for the protocol’s performance and security, potentially heralding significant changes across the entire drone ecosystem.

Community forums saw discussions on complex real-world technical challenges, including RTK GNSS integration, high-speed data communication, specific hardware driver issues, and VTOL aircraft flight stability problems. These activities underscore PX4’s dedication to generating user value not merely through firmware development, but also through a deep understanding and support for practical applications.

Core GitHub Updates (PX4-Autopilot)

Key Merged Pull Requests

• MAVLink Protocol Enhancements and Bug Fixes:
  • fix(mavlink): preserve OpenDroneID system timestamp: Addresses an issue with OpenDroneID system timestamp handling, enhancing compliance with Remote ID regulations and data accuracy.
  • fix(battery): Add invalid sentinel for BatteryStatus values: Adds a sentinel value for validating battery status values, improving the reliability of battery telemetry data. This contributes to consistent battery data processing, alongside the `fix(mavlink): preserve -1 marker in battery fields` (Open) PR.
  • feat(ekf2): Fusion-Control of sensors over MAVLink: Integrates MAVLink-based sensor fusion control into EKF2, significantly improving flexibility in interfacing with external sensors. This is a crucial advancement for EKF2’s scalability.
• Safety and Control System Improvements:
  • Control surface preflight check: Adds a preflight check for control surfaces, bolstering flight safety. This is a vital step in system integrity verification.
  • Bugfix: Let user take over from a degraded failsafe: Fixes a bug allowing users to take over control from a degraded failsafe state, ensuring flexibility for pilot intervention during critical situations.
  • chore(tecs): remove TECS airspeed filter parameters: Removes unnecessary airspeed filter parameters from the TECS (Total Energy Control System), simplifying control logic and improving maintainability.
• Development Infrastructure and Documentation:
  • Hob/uxrce dds stability: Improves the stability of UXRCE DDS (Data Distribution Service), enhancing middleware communication reliability. This is essential for advanced applications requiring high-performance communication.
  • feat(zenoh): add support for configuring zenoh publisher options: Adds support for configuring Zenoh middleware publisher options, expanding the flexibility of the distributed data communication stack.
  • Add secure boot and firmware signature verification documentation: Adds documentation for secure boot and firmware signature verification, an important effort to enhance system security.
  • feat(ci): add certain labels automatically: Adds automatic labeling functionality to the CI/CD pipeline, increasing development workflow efficiency.
• Commander Module Refactoring:
  • refactor(commander): remove unused parameter COM_TAKEOFF_ACT and related PRs (COM_LKDOWN_TKO, COM_ARM_SDCARD, COM_IMB_PROP_ACT, COM_OBC_LOSS_T): Removes unused parameters to reduce the complexity of the Commander module and improve the readability of safety-critical code.

Open Critical and Key Pull Requests/Issues

• Architectural and Core Feature Advancements:
  • [RFC] Reconsideration of the current mavlink implementation: This RFC requests a reconsideration of the current MAVLink implementation. Tagged with `risk:performance`, `risk:security`, and `community:dev-call`, it discusses broad improvements or redesigns for PX4’s MAVLink interface, potentially marking a significant turning point in future development directions.
  • Draft: Static Vision-based target esitmator (Kalman Filter): The draft PR for a static vision-based target estimator (Kalman Filter) is `risk:safety-critical` and spans multiple core scopes including `scope:estimation`, `scope:navigation`, and `scope:control`, holding significant potential for enhancing PX4’s advanced autonomous flight and precise detection capabilities.
  • feat(bootloader): Revive secure-boot with example, docs, and various fixes: This PR reactivates secure boot functionality and provides related documentation and examples. It is a critical feature for fundamentally strengthening the integrity and security of drone systems.
  • feat(navigator): Geofence Aware RTL Direct: Adds Geofence-aware RTL (Return To Launch) functionality, enabling path planning that considers safe zones during emergency return. It is labeled `risk:safety-critical`, indicating high importance.
  • feat(navigator): extend detect and avoid module to follow ASTM F3442 standard: This PR extends the detect and avoid module to comply with the ASTM F3442 standard. It aims for industry standard compliance for safe autonomous drone operations, requiring complex feature implementation and testing.
  • feat(dshot): closed-loop RPM controller using Bidi-DShot feedback: A closed-loop RPM controller utilizing Bidi-DShot feedback will enhance motor control precision, contributing to overall flight performance and efficiency.
• Hardware Support and Drivers:
  • feat(pixi) Add Pixi build system for cross-platform development environment management: The addition of the Pixi build system will standardize cross-platform development environment management, facilitating easier integration of new boards and drivers.
  • feat(drivers/imu): add Analog Devices ADIS1650x IMU Driver and feat(drivers/imu): add Analog Devices ADIS1657x IMU driver: The addition of high-performance IMU drivers for Analog Devices expands PX4’s supported sensor spectrum, with anticipated performance improvements, especially in precision applications.
  • feature(boards): add support for CubeRed and other board addition PRs: Support for new flight controller boards (Matek H743, Corvon V5, CBUnmanned H753-Stamp) expands the diversity of the PX4 ecosystem and offers more choices to users.
  • [Bug] File operations failing (STM32+NuttX+DTCM+IDMA): A critical bug where file operations fail on specific STM32 boards is a significant issue at the hardware abstraction layer, requiring urgent resolution.
  • [Bug] CAN-based RTK GNSS not working on NXP v6X-RT – inconsistent CAN frames, no RTK Fix: A bug preventing CAN-based RTK GNSS from working on NXP v6X-RT boards severely impacts applications relying on high-precision positioning.
• Simulation and Testing:
  • feature(tests): run integration tests with PX4 internal SIH simulation: The move to run integration tests with PX4’s internal SIH (Software-In-the-Loop) simulation is expected to enhance testing efficiency and repeatability. This is linked to the `ros_integration_tests: migrate from Gazebo Classic to SIH` PR, indicating a shift in simulation strategy.
  • Multiple video streams in multi-vehicle simulation: A request for improvement to support multiple video streams in multi-vehicle simulation is essential for building complex autonomous system test environments.

Weekly Dev Call & Community Trends

The Weekly Dev Call held on April 29, 2026, proceeded as a regular team synchronization and community Q&A session; however, no particularly significant agenda items or decisions from the discussions have been publicly specified in the forum to date. This could suggest that a broad consensus on the project’s core direction is being achieved, or that detailed discussions were conducted privately. Nevertheless, activity on the Discourse forum highlighted the active engagement of the PX4 community and the real-world technical challenges being addressed.

Key Discourse Discussions

• PX4 Autopilot Section:
  • Urgent: Quadcopter Tiltrotor VTOL takeoff and then oscillate and drop: An urgent query regarding instability and crashes during takeoff of Quadcopter Tiltrotor VTOL aircraft indicates that optimization and stability assurance of VTOL control algorithms remain crucial areas for research and development.
  • ICM20948 not working on external SPI port: The issue of ICM20948 IMU sensor integration via an external SPI port represents a common challenge in hardware driver and system integration processes, requiring technical support from the community.
  • We Merged a Jetson Carrier Board and an ArduPilot Flight Controller Into One Board. Here Is What We Built: Although this post introduces an ArduPilot-based system, it reflects the community’s high interest in hardware solutions that integrate Jetson with flight controllers. This suggests potential demand for similar integrated solutions combining high-performance onboard computing with PX4.
• Pixhawk Section:
  • How to configure NTRIP and RTK via LoRa on a drone: An inquiry about configuring NTRIP and RTK (Real-Time Kinematic) using LoRa communication on a drone indicates significant interest in building high-precision positioning systems for drones. This is a critical factor for maximizing PX4’s navigation accuracy.
  • Connecting 10BASE-T1S (Pixhawk NXP) directly to onboard computer without media converter: Discussion on how to directly connect the Pixhawk NXP board’s 10BASE-T1S interface to an onboard computer without a media converter demonstrates exploration into utilizing new hardware interfaces for high-bandwidth, low-latency communication. This is essential for integrating complex payloads and computing modules.

Subsystem Trends (MAVLink, MAVSDK, QGC)

MAVLink

MAVLink occupied a central position in PX4 development activities this week. On GitHub, bug fixes such as OpenDroneID timestamp handling and validation of battery status values (PR 27244, PR 27236) were merged, enhancing the protocol’s reliability. Notably, feat(mavlink): Battery_Status_V2 MAVLink stream is expected to strengthen system monitoring capabilities by providing richer battery telemetry data.

The most significant trend is the [RFC] Reconsideration of the current mavlink implementation issue. This RFC raises fundamental questions regarding the performance, security, and scalability of the current MAVLink implementation, suggesting a potential strategic re-evaluation of MAVLink’s role and implementation within the PX4 project. This could lead to broad discussions aimed at evolving the MAVLink communication architecture itself, beyond mere bug fixes or feature additions. Furthermore, the merger of feat(ekf2): Fusion-Control of sensors over MAVLink demonstrates that MAVLink can be extended as a sensor fusion channel for EKF2, proving its expanding scope of application.

MAVSDK

No direct GitHub activity (PRs, Issues) or Discourse discussions regarding MAVSDK were observed in the publicly available data over the past seven days. This could imply that the MAVSDK project itself is in a stabilization phase, or that more resources were concentrated on core Autopilot firmware development. As MAVSDK is an essential interface for high-level application developers, it will be crucial to closely monitor related trends next week.

QGroundControl

Similarly, no significant updates or discussions for QGroundControl (QGC) were reported in its GitHub repository activity or Discourse forum this week. While QGC plays a vital role as PX4’s primary ground control station, this week appears to have been relatively quiet compared to backend firmware development. However, updates to PX4 firmware, such as changes to MAVLink message structures or the addition of new parameters, could trigger future QGC updates, necessitating continuous monitoring.