RoboFacet learning center
Skip the textbook language. Learn what each term means, what it is often confused with, and why it matters when choosing a product—from robot hardware to AI agents.
36 terms
The numbers and interfaces that describe what a physical robot can actually do.
rated payload
The weight a robot can carry at its wrist. Count the gripper, cables and the part—not just the part itself.
Why it matters
Running near the limit can reduce speed and usable reach, so a 10 kg part may need a robot rated well above 10 kg.
working range
The farthest distance the robot arm can reach from its base. It describes an outer limit, not every pose inside it.
Why it matters
Fixtures, approach angles and obstacles can shrink the useful workspace long before the published maximum.
DoF · DOF
The independent movements a robot can control. A typical six-axis arm can position and orient a tool in 3D space.
Why it matters
More axes can reach around obstacles and awkward angles, but they also add cost and control complexity.
repeat positioning accuracy
How closely a robot returns to the same taught point again and again. Think consistency, not whether the point was correct in the first place.
Why it matters
It matters for repetitive assembly and dispensing, but it should not be confused with absolute accuracy.
position accuracy
How close the robot gets to a commanded real-world coordinate without being taught that exact point first.
Why it matters
Offline programming, robot-to-robot handoff and metrology need it; excellent repeatability alone is not enough.
takt time
The time needed to complete one full unit of work, including movement, gripping, waiting and handoff.
Why it matters
A fast robot does not guarantee a fast cell if vision, feeding or machine waiting is the real bottleneck.
EOAT · robot tool
The tool attached to the robot wrist: a gripper, suction cup, welder, screwdriver, camera or something custom.
Why it matters
The arm only moves; the end effector actually touches the job and often decides whether the application works.
cobot
A robot designed with features that can support work near people, such as force limits and easier teaching.
Why it matters
Collaborative does not automatically mean safe without guarding; the tool, speed, payload and task still need a risk assessment.
software development kit
Code libraries, documentation and examples that let developers control or extend a product without starting from zero.
Why it matters
A polished demo with a weak SDK can be far harder to integrate than a less flashy product with open, stable interfaces.
Robot Operating System · ROS2
A robotics software ecosystem that helps sensors, controllers, visualization and applications exchange data. Despite the name, it is not a complete operating system by itself.
Why it matters
ROS 2 support can shorten development, but check the actual drivers, supported version and documentation—not just a logo on the feature list.
Navigation, fleet control and the warehouse systems around moving robots.
autonomous mobile robot
A mobile robot that maps its surroundings and plans routes around people or obstacles instead of following one fixed path.
Why it matters
AMRs suit changing routes, but good navigation does not replace fleet rules, charging plans or process integration.
automated guided vehicle
A guided vehicle that normally follows predefined markers, magnetic tape, reflectors or fixed routes.
Why it matters
AGVs can be simpler and predictable in stable sites; an AMR is not automatically the better choice.
simultaneous localization and mapping
A robot builds or updates a map while estimating where it is on that map—like drawing the room while finding its own dot.
Why it matters
Reflective walls, repetitive aisles, moving crowds or poor lighting can challenge different SLAM methods.
collision avoidance
The robot senses something in its path and slows, stops or finds another route to avoid a collision.
Why it matters
Passing a clean demo is easy; test glass, low objects, blind corners and busy traffic in the real site.
fleet manager · RCS
Software that assigns jobs, prevents traffic jams, manages charging and tracks many mobile robots as one fleet.
Why it matters
A single robot may look excellent while a large fleet performs poorly without good traffic and task scheduling.
warehouse management system
The warehouse system that knows what stock exists, where it is and which orders need to be fulfilled.
Why it matters
Robots can move goods efficiently only when location, inventory and order data are reliable.
warehouse execution system
The execution layer that coordinates people, conveyors, robots and workstations to keep warehouse work flowing.
Why it matters
The WMS decides what should happen; the WES often decides what should happen next, with which resource.
The control, production and integration layer around factory equipment.
programmable logic controller
A rugged industrial controller that reads sensors and controls machines with predictable, repeatable logic.
Why it matters
Robot cells often rely on a PLC for interlocks, sequencing and safety-related coordination with other equipment.
manufacturing execution system
Software that tracks and manages production orders, work in progress, quality and equipment activity on the factory floor.
Why it matters
Automation creates more value when production results and exceptions flow back into the system of record.
computer vision · robot vision
Cameras, lighting, optics and software working together so a machine can locate, inspect or measure something.
Why it matters
The camera is only one part; lighting and real defect samples often matter more than a higher-resolution sensor.
robot integrator
The work of making the robot, tooling, sensors, controls, software and site operate as one complete system.
Why it matters
Buying the robot is often the easy part; integration determines reliability, safety and how quickly production recovers from faults.
virtual commissioning
A digital model connected closely enough to a real machine or process to test, monitor or predict its behavior.
Why it matters
A 3D model alone is not a twin; useful twins need current data, behavior rules and a clear operational purpose.
production stop
Time when equipment should be producing but cannot because of faults, maintenance, changeover or missing material.
Why it matters
A small gain in robot speed can be worthless if recovery is slow and one fault stops the whole line.
Ways to automate screens, documents, APIs and repeatable digital work.
robotic process automation
Software bots that click, type, copy and move data through existing applications much like a person using the screen.
Why it matters
It is useful when no API exists, but interface changes and unexpected cases can break the bot.
workflow · integration automation
Rules that pass information and trigger actions across apps—for example, a form creates a record and sends an approval request.
Why it matters
It usually works through APIs and events, making it more stable than screen-clicking RPA when good connectors exist.
optical character recognition
Technology that turns text in scans, photos or PDFs into characters a computer can search and process.
Why it matters
Reading text is not the same as understanding fields; messy documents still need validation and exception handling.
application programming interface
A defined way for one piece of software to request data or actions from another without clicking through its user interface.
Why it matters
A stable API is usually easier to monitor and maintain than UI automation, but permissions and rate limits still matter.
self-hosted · on-premise
Running software in infrastructure you control instead of relying only on the vendor's hosted cloud service.
Why it matters
You gain control over data and deployment, but also inherit upgrades, backups, security and incident response.
unattended bot
Automation that runs on a schedule or trigger without a person starting and watching every job.
Why it matters
It saves more labor than an attended assistant, but needs credentials, monitoring, retries and a clear exception queue.
The building blocks and safeguards behind model-driven automation.
agentic AI
A model-driven system that can decide steps, use tools and work toward a goal instead of returning only one answer.
Why it matters
More autonomy also creates more ways to fail, so permissions, evaluation and approval boundaries are part of the product.
retrieval-augmented generation
The system finds relevant material from a chosen knowledge source and gives it to the model before generating an answer.
Why it matters
It can ground answers in current company information, but poor retrieval can still produce confident, unsupported responses.
function calling · tools
A model selects a predefined function—such as searching, sending or updating—and supplies structured inputs for software to execute.
Why it matters
This turns chat into action, which means every tool needs narrow permissions, validation and safe failure behavior.
vector store · embedding database
A store that finds items with similar meaning by comparing numerical representations called embeddings.
Why it matters
It helps semantic search and RAG, but chunking, metadata and access control often matter more than database brand.
HITL · human review
A person reviews, approves or corrects an automated decision at a deliberate point in the workflow.
Why it matters
It is most valuable before irreversible, expensive or sensitive actions—not as a vague promise that someone is watching.
evals · benchmark
A repeatable set of examples and scoring rules used to check whether an AI system is useful, accurate and safe enough.
Why it matters
A demo shows what can work once; evaluation shows how often it works and what breaks after a change.
fabrication · confabulation
An AI produces information that sounds plausible but is unsupported, incorrect or invented.
Why it matters
RAG can reduce some errors but cannot guarantee truth; important outputs still need evidence, checks or approval.
