Lesson 6 of 2245 minutes

Actuators and End Effectors

Start with the lesson question, connect the representations, and test the model with evidence.

actuatorsend effectorstorqueforcetransmissionpayloadfeedbacksafety

Learning objectives

  • Explain the responsibilities of sensors, actuators, compute, power, and mechanics.
  • Compare sensors using range, resolution, latency, and failure modes.
  • Select components for a constrained robot task.
Lesson flowHook, model, explanationShow guidance

Inspect the opening phenomenon

Predict what changes, then name the evidence.

Apply in the lab

Name the evidence before reading the answer.

Read only what helps

Then use the lab and recall check.

More when needed

Transcript and resources stay available below.

Course progress

AI & Robotics Foundations · Robot Bodies, Sensors, Actuators, and Compute · Lesson 6

Actuators and End Effectors

In progress

Decision challenge

Observe the phenomenon. Then connect the representations.

Use the opening example to make a prediction, identify evidence, and explain which model supports it.

Can a Robot Grip Harder and Become Less Safe?

Predict why more torque or grip force may reduce safety.

More torque or grip force can reduce safety when contact and payload evidence are missing.

Before

Predict why more torque or grip force may reduce safety.

During

Track commands forward and measured evidence backward.

After

Name three facts the robot must verify before contact.

Reference drawerTranscript, source notes, scripts, and package status stay tucked away until you need them.7 files

Lesson reading

live

45 min

Video script

draft

Transcript fallback

available

courses/ai-robotics/modules/02-robot-bodies-sensors-actuators-compute/lessons/03-actuators-and-end-effectors/video-transcript.md

Action-Chain Design Lab

draft

25 min

Mastery check

live

6 questions / 10 min

Book section:courses/ai-robotics/modules/02-robot-bodies-sensors-actuators-compute/lessons/03-actuators-and-end-effectors/book-section.md
Transcript for accessibility and fallback

# Video Transcript Can a robot grip harder and become less safe? Absolutely. A robot acts through a chain: command, drive, actuator, transmission, joint, end effector, contact—and feedback back to the controller. The actuator converts energy into motion. The transmission trades speed for torque. The end effector touches the task. A high gear ratio may help lift a load, but it slows the joint and can amplify damaging contact. A stronger grip can crush a beaker. Rated payload is not enough; tool mass, center of gravity, inertia, and pose matter. Before the robot moves, name three facts it must verify. Limits, current state, and contact or task-success evidence—with a safe stop if any becomes stale. Continue the full AI Robotics learning path on EduQuest. Subscribe to Humanoid Hub for the next lesson.

Reading lab

Core explanation

Connect the lesson's words, diagrams, graphs, evidence, and equations.

Opening challenge: A gripper closes on a glass beaker. What must the robot know before “close” becomes a safe physical action?

Mental Model: A Robot Acts Through a Chain

An actuator converts supplied energy and a command into physical effort and motion. An end effector is the task-facing device at the end of the chain: a gripper, suction cup, tool, wheel, foot, or other interface with the world.

Robot action chain from command through drive, actuator, transmission, joint, end effector, contact, and feedback

command → drive → actuator → transmission → joint → end effector → contact
   ↑                                                                    ↓
   └──────────── measured position, speed, effort, current, contact ─────┘

The chain matters because a correct high-level command can still fail through insufficient torque, excessive speed, backlash, wrong payload data, a poor tool, uncertain contact, or stale feedback.

Actuator Is a Role, Not a Synonym for Motor

TechnologyConvertsTypical strengthTypical trade-off
Electric motorelectrical energy to rotationprecise, wide control rangeheat, gearing, current limits
Pneumatic cylindercompressed-air pressure to linear motionfast and mechanically simplecompressibility complicates precise force/position
Hydraulic actuatorfluid pressure to force/motionhigh force densitypumps, leaks, maintenance, fluid hazards

ROS 2 control represents robot hardware with command interfaces and state interfaces. A controller may command position, velocity, or effort while reading the resulting state. “Command sent” is therefore not evidence that motion or contact succeeded.

Transmission Changes the Trade-off

A transmission maps actuator motion and effort into joint motion and effort. Ideal mechanical power is approximately conserved:

P=τωP = \tau\omega

where τ\tau is torque and ω\omega is angular speed. Gearing can raise output torque while reducing output speed. Real transmissions also add friction, compliance, backlash, heat, and efficiency losses.

Worked example

A motor can supply 0.40 N\cdotpm0.40\ \text{N·m} through a 20:1 reduction. Ignoring losses, the output torque is about:

0.40×20=8.0 N\cdotpm0.40 \times 20 = 8.0\ \text{N·m}

At 80% efficiency, a better estimate is 6.4 N\cdotpm6.4\ \text{N·m}. The joint rotates about twenty times slower than the motor. This may help lift a load but may make the action too slow—or amplify collision forces if limits and feedback are poorly designed.

End Effector Selection Starts With the Task

Choose from evidence, not appearance:

  • object geometry, material, fragility, surface, and variability;
  • required force, speed, accuracy, and cycle time;
  • payload mass, center of gravity, and inertia;
  • contact sensing and slip detection;
  • failure behavior if power, pressure, vacuum, or communication is lost;
  • cleaning, contamination, maintenance, and human proximity.

Universal Robots requires payload, center of gravity, and inertia information for optimal robot behavior. Picking and releasing an object changes the active payload. A tool that can hold the mass may still overload the arm when the center of gravity is far from the flange.

Safe Action Requires Feedback and Limits

A decision-ready action contract includes command mode, valid range, maximum speed/effort, expected state update rate, stale-state threshold, contact criterion, stop behavior, and reset conditions. ROS 2 hardware lifecycle states separate configuration, activation, deactivation, shutdown, and error handling; movement interfaces should not be treated as always available.

End-effector safety is also an integration problem, not merely a component rating. ISO/TR 20218-1 provides specific guidance for designing and integrating robot-system end effectors, complementing the broader ISO 10218 robot and application safety requirements. This lesson teaches the engineering evidence needed for reasoning; it does not replace a task-specific risk assessment or the applicable standards.

Retrieval Pause

  1. Why can gearing increase torque but not create free power?
  2. Why does a gripper need payload and center-of-gravity information?
  3. What evidence distinguishes “command delivered” from “task succeeded”?

Misconception Checks

“The motor is the actuator, transmission, and joint.” These are separate roles even when packaged together.

“More grip force is always safer.” Excess force can crush, deform, launch, or damage an object.

For example, Robotiq's 2F gripper control interface treats requested position, speed, and force as separate settings. Its force setting limits motor current, and exceeding that limit stops the fingers and triggers object-detection status. That is concrete evidence that “grip harder” is not a complete control strategy: the robot must also observe state and contact outcome.

“Rated payload proves every pose is safe.” Reach, center-of-gravity offset, inertia, acceleration, tool mass, and orientation all matter.

“Position control guarantees safe contact.” Contact also depends on force, compliance, geometry, sensing, and stopping behavior.

Summary

Robots act through a coupled chain. Select the actuator technology, transmission, joint, and end effector together; verify payload and contact assumptions; read state rather than trusting commands; and define limits and safe stopping before physical execution.

Sources and Further Learning

Practice labAction-Chain Design LabOpen this when you are ready to apply the model, collect evidence, and check your explanation.25 min

Objective

Choose and justify an actuator, transmission, and end effector for a fragile-object transfer without requiring hardware.

Setup

Use paper or a spreadsheet. Mission: move a 0.40 kg glass container 0.30 m, place it within 10 mm, and stop safely if grip evidence is lost.

Materials

  • paper and pencil or an accessible spreadsheet;
  • the lesson's robot action-chain visual;
  • a calculator (optional).

Steps

  1. Compare an electric geared motor, pneumatic cylinder, and hydraulic actuator for force, speed, precision, infrastructure, and failure behavior.
  2. Compare a parallel gripper, suction cup, and compliant soft gripper for the object.
  3. Draw the command-to-contact chain and the feedback path.
  4. Create a budget for tool mass, object mass, center-of-gravity offset, desired acceleration, grip force, cycle time, and measurement age.
  5. Evaluate two gear ratios. Explain the torque/speed trade-off and include an efficiency assumption.
  6. Define success evidence: object detected, grip/contact confirmed, motion within limits, object released at target, no stale state.
  7. Define safe responses for object slip, stale feedback, excess current/effort, power loss, and failed release.
  8. Complete an integration gate before accepting the design: identify the lifecycle phase affected (design, commissioning, operation, maintenance, or decommissioning), the person exposed, the foreseeable hazardous event, the protective measure, and the evidence that the measure works. This mirrors the system-level scope described by ISO 10218-2:2025 without claiming standards compliance.

Expected Result

A comparison table, annotated chain, calculations with assumptions, selected design, rejection rationale, acceptance tests, and safe fallback.

Troubleshooting

  • If one option seems best at everything, name its infrastructure and failure costs.
  • If payload is only a mass, add tool mass, center of gravity, inertia, pose, and acceleration.
  • If success means “command sent,” add measured task evidence.

Reflection Questions

Which assumption most threatens the design? What sensor or test would reduce that uncertainty?

Extension Challenge

Change the object to a flexible pouch or porous box. Re-evaluate the end effector, sensing, force limits, failure behavior, and acceptance evidence; explain which earlier assumptions no longer hold.

Accessibility and Safety Fallback

All work may be submitted as structured text. Do not test gripping on people, animals, glass, powered machinery, pressurized systems, or unsecured robots. ISO 10218-2:2025 addresses integration across design, commissioning, operation, maintenance, and decommissioning of industrial robot applications; this paper exercise is educational evidence, not a risk assessment or certification.

Primary reference: ISO, ISO 10218-2:2025 — Industrial robot applications and robot cells, official scope and lifecycle summary; retrieved 2026-08-18. The paid standard text was not reproduced.