What Is a Robot Arm? (And How Does It Move?)

A robot arm is like a human arm, made of metal.

It has joints (like your elbow and wrist). It has links (like your upper arm and forearm). And it has a hand at the end (called the “end effector”) that grabs things.

The hard part? Telling it where to go.


The Parts of a Robot Arm

        [Hand]  ← grabs things
          |
    [Wrist]   ← rotates
          |
   [Elbow]    ← bends
          |
  [Shoulder]  ← connects to the base
          |
       [Base]  ← sits on the floor

Each joint is a motor. Each motor can:

6 joints = 6 degrees of freedom. That’s enough to reach any point in space and point the hand in any direction.


The Two Hard Problems

Robotics engineers spend most of their time on two questions:

Problem 1: Forward Kinematics

“I know the joint angles. Where is the hand?”

This is math. Given each joint angle, calculate the hand position.

It’s like: “If I bend my elbow 90 degrees and lift my shoulder 45 degrees, where is my fingertip?”

The good news: Computers can do this fast. It’s just trigonometry.

Problem 2: Inverse Kinematics

“I know where the hand should be. What joint angles get it there?”

This is the reverse. And it’s much harder.

It’s like: “I want to touch that cup. What angles should my shoulder, elbow, and wrist be?”

Why it’s hard:


How Engineers Solve It

For Simple Arms (2-3 joints)

Use math formulas. Like solving a triangle in high school geometry.

For Complex Arms (6+ joints)

Use computer algorithms that try many angles and pick the best one.

Two common methods:

MethodHow It WorksBest For
AnalyticalMath formula, exact answerSimple arms with standard shapes
NumericalComputer tries angles, step by stepComplex arms, weird shapes

A Simple Example

Imagine a 2-joint arm:

    Hand (x, y)
      |
   Link 2 (length L2)
      |
   Joint 2 (angle θ2)
      |
   Link 1 (length L1)
      |
   Joint 1 (angle θ1)
      |
     Base

Forward Kinematics (angles → position):

x = L1 × cos(θ1) + L2 × cos(θ1 + θ2)
y = L1 × sin(θ1) + L2 × sin(θ1 + θ2)

Inverse Kinematics (position → angles):

cos(θ2) = (x² + y² - L1² - L2²) / (2 × L1 × L2)
θ1 = atan2(y, x) - atan2(L2 × sin(θ2), L1 + L2 × cos(θ2))

Don’t worry about the formulas. The point is: it’s math, not magic.


Real-World Arms You Know

ArmJointsWhat It Does
Factory welding arm6Welds car bodies, same spot every time
Surgical robot (da Vinci)7Lets doctors operate through tiny cuts
Roomba vacuum0 (just wheels)Not really an arm, but still a robot!
Space station arm (Canadarm)7Moves heavy stuff outside the ISS
Humanoid robot (Optimus)40+Walks, grabs, does lots of things

Why This Matters for AI

Old robot arms were dumb. They followed exact paths programmed by engineers.

New robot arms are smart. They use AI to:

The math is still the same. But AI helps the robot figure out the angles on its own — instead of a human programmer doing it.


Key Terms (Simple Definitions)

TermSimple Meaning
JointA motor that bends or spins
LinkA metal bar between two joints
End effectorThe “hand” — gripper, welder, camera, etc.
Degrees of freedomHow many ways the arm can move (usually 6)
Forward kinematicsAngles → position (math, easy)
Inverse kinematicsPosition → angles (hard, needs solving)
WorkspaceThe area the arm can reach
SingularityA position where the math breaks down

Try This at Home

Want to understand robot arms? Build one:

  1. Cheap option: Buy a “robotic arm kit” online ($30-80). It won’t be precise, but you’ll see the joints and links.

  2. Free option: Use a robot simulator online. Search for “robot arm simulator” and play with moving joints.

  3. Code option: Write a simple 2-joint arm in Python. Use the formulas above. Plot the hand position as you change angles.


What’s Next?

Robot arms used to be factory-only. Now they’re:

The math hasn’t changed. But AI, cheaper motors, and better sensors are making arms useful everywhere.

The future: You tell the robot “grab that cup” — and it figures out the rest.


This is part of Robot Basics — our simple guides to how robots work. No math degree needed.