AFirst deep dataset
Motion & Actuation
Joints, drives, transmissions, motors and feedback—normalized at the test-condition level.
Independent robot component database
Robot hardware,
evidence attached.
RCI separates what manufacturers claim, what independent sources verify, what we derive, where sources conflict—and what remains unknown.
Coverage modelFive linked layers
01 / AActuation02 / BPerception03 / CCompute04 / DPower05 / EIntegrationOne robot. Five hardware layers. No isolated spec sheets.
Every category is structured around the questions engineers and procurement teams actually face: performance under stated conditions, interfaces, verified combinations, lifecycle and replacement evidence.
BSource map active
Perception & Sensing
Cameras, LiDAR, radar, inertial and tactile sensing—mapped with calibration and interface evidence.
CSource map active
Compute & Control
Main computers, accelerators and real-time controllers—compared beyond headline TOPS.
DSource map active
Power & Energy
Battery packs, BMS, conversion and charging—tracked with safety, thermal and duty-cycle conditions.
EFoundation layer
Interfaces & Integration
The mechanical, electrical, data, timing, thermal and software edges where components succeed or fail together.
A primary source is not the same thing as independent proof.
RCI labels the status of each claim—not just the reputation of the document that contains it.
Independently verified
Supported by independent testing, a disclosed deployed BOM, or another qualified external record.
Manufacturer claimed
Directly stated by the manufacturer; useful primary evidence, but not independent verification.
RCI derived
Calculated or normalized by RCI with the source values, units and method preserved.
Conflicting
Credible sources disagree. The conflict remains visible until evidence resolves it.
Unknown
No adequate public evidence. Unknown is a valid result, not a blank to fill with inference.
The front edge of robotics is scattered across documents that were never designed to agree.
RCI monitors standards, papers, engineering repositories, patents, filings, government programmes, events and media leads—then records what may be cited, summarized or republished.
43monitored source channels
38primary or first-party channels
5reuse / commentary classes
30fixed AI citation questions
First operating rule
Unknown is a result—not a defect to hide.
RCI will publish missing test conditions, incompatible definitions and unresolved conflicts when those gaps are more useful than a false comparison.
- 01Preserve the original value, unit, test condition and source location.
- 02Keep manufacturer claims separate from independent measurements.
- 03Show conflicts instead of silently choosing the cleaner number.
- 04Publish corrections and material changes with timestamps.
Reference pages first. Original investigations next.
The launch release documents RCI’s boundary and method. Component comparisons publish only after the field-level source audit passes.
RCI 001 · Method notePublished
Why RCI publishes evidence states before product rankings
A manufacturer datasheet can be the best primary source and still be a manufacturer claim. RCI keeps source quality and claim status separate.
RCI 002 · Field guidePublished
Five layers for mapping a robot hardware stack
A practical boundary for actuation, perception, compute, power and the interfaces that bind them into one machine.