Simulating an Obstacle-Avoiding Robot – Robot Logic in Scratch (Senior 3 ICT) | Ganzaa

Robot Logic in Scratch

Lessons

Senior 3 · ICT

Robot Logic in Scratch

5 lessons

Lesson 4 of 5

Simulating an Obstacle-Avoiding Robot

60 min

Never bump, never stop

Robot vacuum cleaners, warehouse carts and planetary rovers share one basic skill: they keep moving through a space full of obstacles without crashing. An obstacle-avoiding robot is the classic first project in robotics because it uses the whole sense–think–act loop with a single sensor. In this lesson you design the algorithm, simulate it in Scratch with a forward-looking sensor, test it, and see how the same logic runs on real hardware.

Definition

Algorithm: a precise, step-by-step set of instructions for solving a problem.

Flowchart: a diagram of an algorithm, using ovals for start and end, rectangles for actions, diamonds for decisions and arrows for the order.

Detection range: the distance at which the sensor first notices an obstacle.

Reaction distance: how far the robot travels between a sensor reading and the end of its response (stopping or turning).

Step 1: the algorithm

Start at the entrance, facing forward.
Read the distance sensor.
If an obstacle is closer than the threshold: back off a little, then turn by a random angle.
Otherwise: move forward a small step.
Go back to step 2, forever.

Flowchart drawn by Ganzaa.

Why a random turn? If the robot always turned exactly 90° right, it could bounce between the same two walls forever, or circle one corner. A random angle sends it in new directions, so it explores the whole space.

Step 2: a forward-looking sensor

The distance to block from the last lesson sees in every direction. A real ultrasonic sensor looks only forward. To imitate it, draw a thin cyan sensor beam sticking out in front of the robot in its costume, 40 pixels long. Draw the arena walls and obstacles on the backdrop in brown. Then color [cyan] is touching [brown]? is true only when something is within 40 pixels ahead: a detection range of 40.

Diagram drawn by Ganzaa. Keep the beam narrow and the same colour everywhere; nothing else on the stage may be cyan.

Step 3: the script

Block diagram drawn by Ganzaa in Scratch 3 colours.

Algorithm step	Blocks
Start at the entrance	go to x: (−180) y: (−120); point in direction (90)
Read the sensor	<color [cyan] is touching [brown]?>, inside the if
Obstacle: back off and turn randomly	move (−5) steps; turn right (pick random (90) to (150)) degrees
Count the turns	change [turns] by (1)
Clear: move forward	move (3) steps
Repeat	forever

Worked example 1: will it stop in time?

The robot's front is 100 pixels from a wall and it moves 3 pixels per frame. Its beam reaches 40 pixels ahead, so it detects the wall when its front is 40 pixels away, after travelling $100 - 40 = 60$ pixels:

$\frac{60}{3} = 20 \text{ frames} \approx \frac{20}{30} \approx 0.67 \text{ s}$

It then backs off 5 pixels and turns, never touching the wall. Now suppose a pupil sets the speed to 50 pixels per frame “to make it faster”. In one frame the beam tip can pass from in front of a thin wall to beyond it, and the robot itself can end up inside the wall. A safe rule for the simulation: the step must be much smaller than the beam length.

Worked example 2: choosing a threshold for a real robot

A real robot drives at 0.3 m/s. Its ultrasonic sensor gives 10 readings per second, and its motors need 0.2 s to stop.

Between two readings it travels $0.3 \times \frac{1}{10} = 0.03$ m.
While stopping it travels up to $0.3 \times 0.2 = 0.06$ m.
Reaction distance: $0.03 + 0.06 = 0.09$ m.

The threshold must be more than 0.09 m. Sensors are not perfect, so engineers add a safety margin: a threshold of 0.15 m (15 cm) is a sensible choice. If the robot is made twice as fast, the reaction distance doubles to 0.18 m and the threshold must grow too.

The same robot in hardware

Job	Simulation	Real robot (for example EV3 or a micro:bit buggy)
Sense	<color [cyan] is touching [brown]?>	Ultrasonic sensor: (distance) < (15)
Back off	move (−5) steps	Both motors backwards for a short time
Turn	turn right (pick random (90) to (150)) degrees	One motor forwards and the other backwards for a random time
Move forward	move (3) steps	Both motors forwards

A wheeled robot turns by driving its two wheels differently: if the left wheel goes forward and the right one backward, it spins to the right on the spot.

Warning

Why the simulated robot gets stuck, and fixes

The beam starts inside a wall: the robot spins on the spot forever. Start it in open space.
Narrow gaps: the robot's body is wider than its thin beam, so a corner can catch its side. Use a wider beam, or add side beams (next lesson).
Colour clashes: a brown picture on the backdrop counts as a wall. Keep walls and decorations in different colours.
The turn sends the beam straight into another wall: in a tight corner it may need several turns. Counting turns helps you spot it.

Blue Butterfly: a 3D animation of how an ultrasonic distance sensor works, the sensor most obstacle-avoiding robots use. Watch how the echo time gives the distance.

✦

Hands-on: build and test the avoider

In Robot_Sim, paint an arena backdrop with brown walls and at least four brown obstacles of different sizes. Leave gaps wide enough for the robot.
Add a new robot costume with a 40-pixel cyan beam in front. Build the avoider script with the Pen, so you can see the path.
Run it for 60 seconds. Record the number of turns and whether it ever touched a wall. Did it visit every part of the arena?
Test three beam lengths (20, 40 and 60 pixels) and three speeds (2, 3 and 6). Record the results in a table: crashes and turns. Which combination works best? Why?
Change the turn to a fixed 90°. Watch the path. Does the robot get trapped? Compare with the random turn.
Add a timer rule: if turns increases by more than 5 in 3 seconds, turn 180° to escape a corner. Test it.
Write the algorithm for your final version as a numbered list and a flowchart.

Practice

Define an algorithm and a reaction distance.
Draw the flowchart of the obstacle-avoiding algorithm using the correct symbols.
Why does the robot turn by a random angle instead of always 90°?
Explain how a cyan beam in the costume imitates a forward-looking distance sensor.
A robot moves 4 pixels per frame and detects walls 60 pixels ahead. Its front starts 180 pixels from a wall. After how many frames does it detect the wall?
A real robot drives at 0.5 m/s, takes 5 readings per second and needs 0.3 s to stop. Find its reaction distance and suggest a threshold.
A simulated robot sometimes passes straight through a thin obstacle. Analyse two possible causes.
A team wants its robot cleaner to “always turn right 90°” because it is easier to predict. Evaluate this design for a room with corners and furniture.

Show answersHide answers

Algorithm: a precise step-by-step set of instructions for a problem. Reaction distance: how far the robot travels between a reading and the end of its response.
Oval Start → rectangle “Read distance sensor” → diamond “Obstacle closer than threshold?” → Yes: rectangle “Back off; turn randomly”; No: rectangle “Move forward a step” → both arrows return to “Read distance sensor”.
A fixed turn can trap it in repeating paths (between two walls or in a corner); a random angle sends it in new directions so it explores and escapes.
The beam is part of the robot's costume, sticking out in front. The test color [cyan] is touching [brown]? is true only when a brown wall lies within the beam's length ahead, as a forward sensor would detect it.
It detects when its front is 60 pixels away: after $180 - 60 = 120$ pixels, $\frac{120}{4} = 30$ frames.
Between readings: $0.5 \times 0.2 = 0.1$ m; stopping: $0.5 \times 0.3 = 0.15$ m; reaction distance 0.25 m. A threshold of about 0.35–0.40 m adds a safety margin.
The step is too large (high speed), so the beam jumps past the thin obstacle in one frame; or the obstacle's colour is not exactly the brown in the sensing block, so it is never detected. (Also: the beam is too short or too thin to touch it.)
Predictable, but poor in practice: in a corner it can turn between two walls again and again, and in rooms with furniture it tends to repeat the same loops and miss areas. Random turns, or turning towards the more open side with two sensors, clean more of the room.

Sources and credits

Obstacle avoidance and Differential wheeled robot, Wikipedia contributors, Wikipedia, licensed under CC BY-SA 4.0. Adapted by Ganzaa.
Scratch is a project of the Scratch Foundation; block names and colours follow Scratch 3.
Video: How Ultrasonic Sensor Works? (3D animated), Blue Butterfly, embedded from YouTube.
The flowchart, arena diagram, block diagram, tables, worked examples, hands-on task and practice items were written by Ganzaa.

This lesson adapts CC BY-SA material and is shared under CC BY-SA 4.0.

1Never bump, never stop

2Step 1: the algorithm

3Step 2: a forward-looking sensor

4Step 3: the script

5The same robot in hardware

6Hands-on: build and test the avoider