Unit 01 · lesson

Safe Work Is Part of the Architecture

Safety is often taught as a list taped to a wall. In robotics, that is not enough.

A robot can contain stored electrical energy, moving mechanisms, pinch points, sharp edges, rotating shafts, hot components, batteries, software faults, and unexpected motion. A system can be "working correctly" and still be unsafe.

The better model is to treat safety as another engineering layer.

Hazard, control, evidence

A hazard is a condition that can cause harm. A control reduces the chance or severity of that harm. Evidence shows whether the control is actually in place.

Consider an unexpected motor start.

Item	Example
Hazard	wheel or arm moves while hands are near it
Control	power removed during mechanical changes
Control	physical support prevents stored motion
Control	enable state is deliberate and visible
Evidence	power source disconnected
Evidence	test occurs in a clear zone
Stop condition	motion differs from expected behavior

Notice what is missing: "be careful."

"Be careful" is not an engineering control.

Default course operating model

This course is simulation-first. You can complete the required reasoning with diagrams, calculations, recorded data, browser tools, or robot simulators. Physical hardware is an optional extension when an approved environment and qualified supervision are available.

For any physical robot activity:

make mechanical changes only in a de-energized state;
do not bypass guards, interlocks, current protection, or emergency controls;
secure mechanisms that can fall, swing, or release stored energy;
keep body parts outside the mechanism's motion envelope during powered tests;
test one bounded behavior at a time;
stop immediately if motion, sound, heat, smell, current draw, or structure differs from the expected condition.

The point is not fear. The point is controlled evidence.

Build a test boundary

Suppose a robot arm should rotate 15 degrees.

A weak test plan says:

Turn on the arm and see if it works.

A stronger plan says:

Expected motion: +15° around shoulder joint
Starting state: mechanism clear, load removed
Maximum command: low test output
Observation: encoder angle + visual direction
Stop if: wrong direction, binding, unexpected current, or motion exceeds boundary
Evidence saved: commanded angle, measured angle, result

That small plan changes the experiment. It tells you what success looks like before motion begins.

Add safety to your system map

Return to your Unit 1 robot diagram.

Add:

one energy source;
one motion hazard;
one human control point;
one stop condition;
one piece of evidence that would prove the system is in a safe test state.

That map is the first page of your robotics systems dossier. You will keep improving it throughout the course.