PUMA Project: ROS 2 Integration & Quadruped Autonomy
- Jan 19
- 2 min read
Updated: Jan 21
Overview
Maxwell Robotics delivered an end-to-end integration project to connect a commercial quadruped robot, the DeepRobotics Lynx M20 Pro, into a larger enterprise robot fleet ecosystem. The work was carried out in collaboration with a leading IT consulting firm for an enterprise customer, with the goal of turning a standalone quadruped into a fleet-ready, remotely operable, autonomous-capable system.
At the core: a ROS 2 integration layer that exposes robot control and sensor streams in a standardized interface, enabling higher-level autonomy, teleoperation tooling, and fleet-side connectivity.
The Challenge
Enterprise fleet systems expect robots to present consistent capabilities:
A reliable control interface (motion, modes, safety states)
Standardized sensor access (e.g., camera feeds, range data, status telemetry)
A clean integration point for autonomy behaviors
Operational tooling for remote monitoring and teleoperation
Quadrupeds are powerful mobility platforms, but they typically ship with proprietary APIs and ecosystem assumptions. The project objective was to bridge that gap and make the quadruped behave like a first-class member of a multi-robot fleet.
What Maxwell Robotics Built
ROS 2 Interface Layer (Robot + Sensors)
A ROS 2 integration layer was developed to expose key robot capabilities through well-defined interfaces:
Robot control and motion commands
Robot state, health, and telemetry
Sensor data streams (e.g., cameras and additional payload sensors, depending on deployment)
A structured integration point for higher-level applications
This created a stable foundation for autonomy and fleet integration, while keeping the system modular and extensible.
Onboard Autonomy Stack (Built on Top of the ROS 2 Layer)
On top of the interface layer, an onboard autonomy stack was implemented to support:
Teleoperation for reliable remote driving and intervention
Waypoint following for repeatable autonomous patrol behaviors
Operational readiness features such as clear state feedback and control handover paths (autonomy ↔ operator)
Comments