What is a collaborative robot (cobot)?

A collaborative robot (cobot) is a multi-joint robotic arm built to work safely alongside people in a shared workspace—not a caged industrial arm. Learn payload, reach, cycle time, EOAT, and TCP here, then follow the how-to guides.

Roooll what-is-cobot guide: collaborative robotic arm shared workspace with payload, reach, and EOAT callouts for buyers

Quick answer

Collaborative robot (cobot): a multi-joint robotic arm designed to work in shared spaces, using speed/force/monitoring strategies to keep risk within your plan

“Collaborative” does not mean zero safety work: you still need a cell risk assessment, zoning, speed/human-distance assumptions, and an I/O scope for acceptance

Start from five selection terms: payload, reach, cycle time, EOAT, and TCP—specs alone do not guarantee you clear the worst pose loop

Match the station first: the same cobot class can require different EOAT and TCP depending on the process (assembly, gluing, loading, inspection)

Early in an automation project, spec sheets throw around collaborative robot, cobot, and robotic arm—they usually mean the same class of equipment: a multi-joint, programmable collaborative robotic arm that picks, places, assembles, or inspects on a line or in a lab.

The difference from a traditional industrial arm is not “does it move?” but whether people and the arm can share the same workspace. Cobots are typically compact, faster to deploy, and designed with force limiting or speed control so nearby tending, loading, or inspection stays within your risk plan. Industrial robot safety requirements are commonly framed in ISO 10218; collaborative applications also reference ISO/TS 15066 (including power-and-force limiting discussion). Collaborative does not mean zero safety work—naming a standard does not replace a cell risk assessment. You still define zones, speeds, and person distance. Scope safety and I/O before you quote → Safety & I/O guide

Where cobots usually land

Use this industry → typical station tasks map to find your starting point (most industries run more than one cell—the table lists common first deployments):

IndustryTypical cobot robotic arm tasks / stations
Smart manufacturing / general industryAssembly, screw driving, welding assist, vision inspection, line-side feeding, changeover support
Automotive & partsSub-assembly, machine tending, gauge sampling, adhesive / dispensing assist
Electronics & PCBBoard gluing, AOI re-check, insert / placement assist, test load-unload
Semiconductor & electronics assemblyTest load-unload, carrier / burn-in rack transfer, re-check repeat handling, limited collab in controlled areas (match cleanroom / process grade)
Machining & metalCNC / mill-turn tending, deburr assist, press pick-place, gauge checks
Rubber, plastics & injection moldingInjection mold tending, degate / trim assist, insert placement, in-mold labeling support
New energy & batteriesCell / module transfer, leak-test tending, PACK line assembly assist, EOL test stations
Logistics & warehousingPalletizing / depalletizing, sort assist, unitizing, end-of-line case pack
E-commerce & fulfillmentParcel sort assist, order kitting, returns inspection, pre-label / weigh pick-place
Food & packaging (production)Case pack, label assist, light sort (match food / cleanroom grade to your spec)
Personal care & FMCGFill / cap assist, gift-box packing, multi-SKU changeover pick-place
Medical & clinical labAnalyzer loading, sample / plate transfer, repeatable bench aliquoting, traceable motion
Hospital rehab & therapyRehab assist paths, therapy stroke replay (scope to clinical workflow and local rules)
Pharma & biotechPlate / carrier transfer, clean-bench repeat motion, batch changeover by program
Quality & metrologyCMM / gauge sampling, repeat measure points, SPC inspection load-unload
Retail & front-of-house serviceRecipe pours, dessert / bar prep, smart-dining front-station cadence
Education & researchCurriculum demos, competition-season roll-in, programmable research pick-place
Consumer goods & 3C packagingCarton pack, case seal assist, multi-SKU changeover pick-place

→ Roooll applications by industry: Smart Manufacturing · Medical & Lab · Retail & Service · Education & Research

Evidence: don’t confuse “arm price” with “project price”

Cost itemPublished range / what it impliesSource
Robot hardware (system hardware)~25–40%AMD Machines
Integration engineering~30–50%AMD Machines
EOAT (end-of-arm tooling)$2k–$40k+Robolist TCO model
Integration engineering (scale)$8k–$60kRobolist TCO model

Collaborative safety framework (what to name in your RFQ)

FrameworkWhat you should require in the RFQSource
ISO 10218Robot & integration safety frame: write validation scope and responsibility boundariesiso.org
ISO/TS 15066Collaborative operation supplement: used for contact power/force assumptionsiso.org

Roooll r-Series spans r-Lite (standard, integrated, mobile S/L), r-Core, r-Reach, r-Max (16 kg / 20 kg), and r-Ultra—from compact cells and line-ready defaults to long-reach transfer and heavy payload. Learn the terms below first—then the comparison table takes minutes, not meetings.

Five terms to know before you shortlist

Payload — how much weight the arm can carry in stable operation

Rated payload in the catalog is for continuous work; peak/instant payload applies only to brief moves. Count part + gripper/vacuum/adapter + external cabling—not part weight alone. Under-estimating payload leads to drops, protective stops, and takt drift on the floor.

→ How to calculate and leave margin: Payload guide

Reach (working radius) — can you get there in the pose you actually run?

Catalog reach is a manufacturer reference under stated conditions—it answers how far the arm is rated to extend, not whether your station is covered at your real pick/place angle. Account for TCP (below), EOAT length, cable dress, and obstacles together.

→ How to verify catalog reach vs critical poses: Reach guide

Cycle time — how many seconds one full loop really takes

Catalog max TCP speed is not your stable cycle time. Add pick, transfer, place, gripper open/close, vacuum build/vent, and inspection waits line by line, then pilot on real parts.

→ How to estimate loop time vs line takt: Cycle time guide

EOAT (end of arm tooling) — the arm’s “hand”

Grippers, vacuum cups, custom nests, and adapter plates—the last interface between the task and the robot flange. Pick the wrong path (grip when you should vacuum, clamp when you should locate) and both payload and takt math break.

→ Gripper vs vacuum vs fixture paths: End-effector guide

TCP (tool center point) — the reference point for motion and measurement

Programs and paths are defined around the TCP, not the flange center. Change the end effector and the TCP moves; when you ask “is reach enough?”, measure base to TCP, not base to flange only.

Common misunderstandings (before you quote)

“Collaborative = no safety talk.” You still define zones, e-stop, and I/O → safety guide (link above)

“Part weight only.” EOAT and cabling missing → payload guide

“Catalog reach is enough.” Critical pose and retract path never checked → reach guide

“Demo runs = takt OK.” Production takt must include waits and the slowest loop → cycle time guide

→ More selection mistakes: Selection mistakes guide

Common questions

Does “collaborative” always mean we can skip fencing?
Not necessarily. Collaborative operation depends on speed/force/monitoring strategies and your human-distance assumptions. Whether fencing is needed is decided by your risk assessment and task boundaries—not by a brochure label.
How should I interpret “collaborative” in a quote?
Treat it as a configuration type, not an acceptance criterion. Your RFQ/SOW should state the key pose speed, human-distance strategy, EOAT contact assumptions, and the delivery boundary for e-stop and I/O.
What parameters should I look at first?
Start with worst-case: payload (incl. EOAT), reach (by TCP and pose), cycle time (sum segments), EOAT path, and TCP as the programming and measurement reference. Learn the five terms first, then deep-dive the matching how-to.
If people are very close, can we still use a cobot?
Often yes—if you land the collaboration strategy in the cell. Write human paths, e-stop accessibility, and safety zone rules into the I/O and risk assessment, then validate one worst-case run and document pass/fail criteria.

Next steps

Write one page using the five terms (worst-case payload, critical TCP points, target takt, EOAT path preference)

Scan the full r-Series lineup specs, then pick 2–3 variants from r-Lite, r-Core, r-Reach, r-Max, and r-Ultra in one comparison table → Side-by-Side Comparison

Narrow by application first? → Product Advisor

Have a station sketch or video? → Contact us

→ How to compare models before you buy: Side-by-Side Comparison guide

Share article

New possibilities for your next cobot deployment.

Explore new ways to move your decision forward—with clarity, confidence, and less second-guessing. You don't need every detail settled before you loop in procurement or engineering. When the guides have pointed the way, the paths below help you take the next step together.