Platforms

We develop software that can be carried across robot types, and we choose the platform to suit the research question.

Robots shown on this page are made by their respective manufacturers. Axinome does not manufacture, sell or distribute robot hardware, and does not claim partnership or distribution agreements with these manufacturers. Images are labelled as manufacturer images or as simulation renders.

Rainbow Robotics RBQ-10 quadruped robot shown standing, with a LiDAR sensor mounted on its back. — Manufacturer image Rainbow Robotics RBQ-10. Third-party hardware, shown as a reference research platform. Image: Rainbow Robotics.

Reference platform: Rainbow Robotics RBQ-10

The RBQ-10 is a quadruped robot made by Rainbow Robotics in Korea for mobility and missions in outdoor and unstructured environments, including slopes and stairs. The manufacturer offers it for inspection, patrol and observation work and as a research platform.

We use it as a reference platform because it combines legged mobility with an open software interface and room for additional sensors, which are the conditions our perception, mission and evaluation work depends on.

Published by the manufacturer

Weight: 42 kg
Maximum payload: 15 kg
Ingress protection: IP54
Software interfaces: C/C++ SDK and ROS 2

Source: Rainbow Robotics product page and RBQ documentation. Check the manufacturer for current specifications.

Where Axinome’s work would sit

Rainbow Robotics designs and manufactures the RBQ-10, including its locomotion control. Axinome’s role is limited to the software, integration and evaluation work described here.

Software: Task and mission logic built on the manufacturer’s SDK and ROS 2 interface, kept separate from platform-specific code.
Sensing: Integration and calibration of additional cameras, depth sensors or LiDAR, and fusion with the robot’s own sensing.
Mission integration: Routes, inspection points and operator tools that connect the robot to a site’s workflow and data systems.
Evaluation: Structured trials that measure navigation reliability, intervention frequency and task completion in representative conditions.

Robot types and the research they support

Different questions call for different bodies. The same structure of interfaces, simulation counterparts and evaluation applies to each.

Quadrupeds

Research uses: Mobility on stairs, slopes and uneven ground; inspection rounds; sensor research in outdoor and industrial spaces.
Typical Axinome work: Mission integration, payload and sensor mounting, localisation in changing environments, and evaluation of gait-level behaviour against task requirements.

Reference platform: Rainbow Robotics RBQ-10

Wheeled robots

Research uses: Autonomous navigation, indoor logistics and long-duration operation on flat floors.
Typical Axinome work: Mapping and localisation, path planning among people and obstacles, docking and charging routines, and fleet coordination.

Simulation render of the Hello Robot Stretch 3, a wheeled research robot with a vertical lift and a telescoping arm. — Simulation render Hello Robot Stretch 3, a wheeled research platform. Simulation render by Axinome of the open MuJoCo Menagerie model; not a photograph.

Manipulators

Research uses: Grasping, placement, assembly steps and other tabletop or workcell tasks.
Typical Axinome work: Perception-driven grasp selection, motion planning, task sequencing and recovery, and operator interfaces for supervised work.

Simulation render of a ROBOTIS OMY collaborative arm on a laboratory table with test objects. — Simulation render ROBOTIS OMY collaborative arm in an illustrative tabletop workcell. Simulation render by Axinome; not a photograph of an Axinome facility.

Mobile manipulators and humanoids

Research uses: Embodied task research that combines moving through a space with handling objects in it.
Typical Axinome work: Coordinating base and arm motion, learning from demonstrations, and evaluating learned behaviours across repeated trials.

Simulation render of the Rainbow Robotics RB-Y1, a wheeled mobile manipulator with two arms and a torso. — Simulation render Rainbow Robotics RB-Y1 mobile manipulator. Simulation render by Axinome of the open MuJoCo Menagerie model; not a photograph.

Sensors and payloads

Much of the practical work on any platform is in its sensors: mounting, calibration, time synchronisation, coordinate frames and drivers. Errors here show up later as perception or navigation failures, so we treat sensor integration as part of the experiment, with documented calibration and recorded configurations.

Typical payloads include RGB and depth cameras, 3D LiDAR, thermal and pan-tilt-zoom cameras for inspection, and grippers or tools for manipulation.

Simulation render of an Intel RealSense D435i depth camera seen from the front, showing its lenses. — Simulation render Intel RealSense D435i depth camera, an example of the sensors we integrate. Simulation render by Axinome.

Carrying software across platforms

Task software talks to a robot through a thin layer of interfaces for motion, sensing and state. Platform-specific code lives behind that layer, together with a simulation counterpart of the same robot. Moving to a new platform means implementing that layer again and repeating the evaluation, not rewriting the task logic.

Working with a specific robot?

Tell us which platform you use or are considering, and what you want it to do.

Discuss a project