Plugin-first
Platform behavior, integrations, and operator views should be extendable through explicit plugin contracts.
Plugin-first operations architecture for heterogeneous autonomous systems.
Status: Early-stage architecture and open-source foundation
NEXUS is a public open-source foundation for a modular robotics operations platform. It is designed around heterogeneous autonomous systems, where different robots, simulators, middleware stacks, and external services can be connected through explicit adapters instead of being hard-coded into one runtime.
The project focuses on architecture, module boundaries, capability-driven control, plugin contracts, and operator-facing system design. It is intended to grow into a robot-agnostic operations layer while staying independent from any single robotics middleware.
NEXUS is currently an early-stage architecture and open-source project foundation. This page presents intended system design, module boundaries, and integration strategy rather than a complete production implementation.
The platform is organized around a normalized operations model. Robots, simulators, and external systems connect through adapters; adapters expose state and capabilities; the mission/task layer works against those capabilities; plugins extend behavior and UI surfaces without tying the platform to one robot type or middleware stack.
Platform behavior, integrations, and operator views should be extendable through explicit plugin contracts.
The core platform should reason about normalized robots and capabilities, not one specific vehicle, arm, drone, or middleware package.
Robot-specific details belong behind adapters that translate external systems into stable NEXUS contracts.
Tasks should request abilities such as navigation, inspection, streaming, or manipulation instead of calling platform-specific commands directly.
ROS, MQTT, MAVLink, simulators, and custom APIs are integration choices, not assumptions baked into the platform core.
The operator interface should support modular panels and tools that reflect available plugins, robot types, and mission contexts.
Defines how new robots, simulators, middleware bridges, and external systems connect to NEXUS.
Tracks declared robot capabilities and exposes them to task logic, UI modules, and plugins.
Provides a controlled extension surface for platform features without modifying the core system.
Coordinates high-level work using normalized state and capabilities instead of robot-specific control paths.
Represents identity, pose, health, telemetry, mode, and availability in a consistent platform model.
Provides the UI shell for supervision, commands, status, diagnostics, and plugin-provided panels.
Hosts bridge concepts for ROS, ROS 2, MQTT, MAVLink, simulators, and custom infrastructure.
Supports debugging and operations through events, logs, state history, and integration diagnostics.
NEXUS treats robotics middleware as an integration boundary. ROS, ROS 2, MQTT, MAVLink, simulators, and custom APIs can all be valuable, but none of them should define the core platform identity. The platform core should depend on stable adapter contracts, normalized robot state, and declared capabilities.
This design keeps the operations layer portable across heterogeneous systems. A UAV using MAVLink, a ROS 2 ground robot, a simulator, and a custom external service should be able to appear inside the same operations model when their adapters expose compatible state and capability contracts.
Define the first shared model for robot identity, state, telemetry, health, commands, and capabilities.
Build the adapter contract that external systems use to connect into the platform.
Add the first simulated robot adapter to validate the model without depending on hardware.
Create initial ROS, MQTT, and MAVLink bridge prototypes around the same adapter boundary.
Build a minimal web/operator console for robot state, capability inspection, and basic commands.
Publish small example plugins that demonstrate extension points and expected project structure.
The NEXUS open-source foundation is available on GitHub.