Teach the thinking
behind the blocks.

Start here: the unit at a glance

What learners will understand

• Robots execute precise instructions in order; they do not infer the intended goal.
• Heading changes the meaning of forward. A turn does not travel to another tile.
• Loops can reduce written code while performing the same number of actions.
• Sensor data informs conditions; IF / ELSE chooses one branch for each decision.
• A useful test compares a prediction with observations and leads to a reasoned change.
• A model leaves things out. A real robot has imperfect motion and sensor readings.

Nine-week course map

Choose the right pace

Use the nine-week sequence for classroom instruction and evidence collection. For a 45-60 minute introduction, sample the first three weekly games, demonstrate the sensor, and postpone independent assessment of nested decisions. For younger beginners, split each lesson into two 25-minute sessions. Give additional time for access needs rather than treating speed as achievement.

Before the first lesson

• Open the course and run a short test on the teaching device. Confirm all nine game links, the robot scene, and the relevant program or lab controls are visible.
• Decide whether learners will use the public site or the offline HTML. Progress is stored in that browser on that device, not in a teacher gradebook; a different browser or site origin may have different progress.
• Print the relevant route map and worksheet. Provide an equivalent 2D or tactile route when a 3D view is not a useful access route for a learner.
• Teach the roles: the driver controls the device; the navigator predicts and traces. Both must be able to explain the program.
• Plan an individual check. A partner’s successful program and game awards alone do not demonstrate each learner’s understanding.

A consistent teaching routine

New York State standards and evidence

NYSED organizes these standards into K-1, 2-3, 4-6, 7-8, and 9-12 bands. Expectations are organized for achievement by the end of the relevant band. Grade 4-5 teaching can build toward the 4-6 expectations; this guide does not change the end-of-band expectation. NYSED: Computer Science & Digital Literacy FAQ

Use the alignment accurately

• The 4-6.CT.8 clarification permits working with loops and conditionals separately. Combining them in the sorting and Mars games is an additional challenge, not a requirement inferred from that standard alone.
• Using a provided 3D simulation is not the same as developing a computational model. Use the Scratch bridge before claiming evidence toward 4-6.CT.1.
• For 4-6.NSD.2, include input, processing, storage, and output in the learner’s systems model. Merely naming a front sensor is insufficient.
• The unit emphasizes Computational Thinking and selected Networks & System Design and Digital Literacy expectations. Other standards, including the wider Cybersecurity curriculum, need additional instruction.

Teach the nine weekly games accurately

Troubleshooting

• A required branch or loop is empty: select its insertion slot and add an action.
• A turn seems wrong: check the written heading and use a physical arrow; camera view is independent.
• A game goal is incomplete: read the challenge and trace. Reaching a destination alone may omit pickup, docking, unloading, or required test evidence.
• A learner cannot interpret the 3D scene: provide the 2D map, accessible state text, counter, and spoken explanation. Record reasoning independently of input method.
• Work seems missing: check device, browser, and public/offline origin. Keep important learning evidence in downloaded worksheets or an approved teacher record.

Week 1: Parcel Post

50-minute lesson sequence

Model program and expected behavior

move; move; pick up; move; move; deliver

Six actions: move twice, pick up, move twice, deliver. The parcel remains on the tile until pickup. Delivery requires both the correct location and carried cargo.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Use six illustrated action cards and a parcel counter. Ask learners to arrange the cards and explain the pickup position. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Compare a route plan grouped as travel / collect / travel / deliver with the same six individual actions. Explain why the order cannot be arbitrary.

Independent exit check

Prompt: Which instruction changes the parcel from waiting to being carried?

Teacher key: Pick up at C3 R3. Moving across the parcel does not collect it automatically.

Connect to Scratch or physical robotics

In Scratch, use a cargo variable and touching-destination condition. A physical gripper would need an explicit pickup action and a successful-grip check.

Week 2: Harbor Pilot

50-minute lesson sequence

Model program and expected behavior

move x2; right; move x2; left; move x2; dock

The boat travels north twice, turns right to east, travels twice, turns left to north, travels twice, and docks. Nine actions visit both beacons in order and satisfy the north-facing docking requirement.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: north. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Provide an arrow counter and mark the two numbered beacons. Use cardinal-direction cards independent of camera angle. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Turn the boat right then left at the start. Explain the same endpoint and the two extra actions.

Independent exit check

Prompt: Why can a boat reach the dock but still fail the docking check?

Teacher key: It must also face north and have visited both beacons in order. Heading and position are different.

Connect to Scratch or physical robotics

Scratch direction 0 points up; this simulator internally uses east/south/west/north indices. Use visual arrows when transferring rather than copying numeric heading values.

Week 3: Loopworks Factory

50-minute lesson sequence

Model program and expected behavior

repeat 4 { move; process station }

Repeat 4 containing Move and Process runs eight primitive actions and processes each of four stations exactly once. Three written blocks express the routine. Processing a station twice fails; moving without processing leaves it unfinished.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Start with one station, then repeat a physical two-card pattern. Keep the loop count visible. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Compare repeat 3 plus one move/process pair with repeat 4. Explain the same executed actions and different written structure.

Independent exit check

Prompt: How many actions run when two actions are inside repeat 4?

Teacher key: Eight actions: four moves and four processing actions. A shorter written program can perform the same work.

Connect to Scratch or physical robotics

In Scratch, a repeat block can contain movement plus a costume, sound, or variable update. Preserve the order of the whole repeated routine.

Week 4: Cave Sentinel

50-minute lesson sequence

Model program and expected behavior

repeat 4 { scan; if clear { move } }; confirm stop

The scout scans and moves twice to C3 R3. The next two scans are blocked by the barrier at C4 R3, so IF skips Move. Confirm stop records the decision using the blocked reading. A scan is required at the current pose before a move or IF check.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Use clear/blocked cards and a tangible barrier. Offer the outer Repeat frame while learners choose scan and branch placement. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Explain the stored reading and why a changed position or heading would make it stale. Recreate the probe in Scratch with explicit values.

Independent exit check

Prompt: What should happen after the scout scans the barrier?

Teacher key: The scan is blocked, so IF clear skips its move. Confirm stop records the decision at the safe tile.

Connect to Scratch or physical robotics

The simulation has ideal map-based sensing. Real distance sensors can be noisy, and a safe classroom robot needs a tested threshold and stop behavior.

Week 5: Color Cargo Sorter

50-minute lesson sequence

Model program and expected behavior

repeat 4 { scan; if RED { send left } else { send right } }

Four scans and four bin actions sort RED, BLUE, BLUE, RED into left, right, right, left. A repeat loop and IF RED / ELSE are required. Every new parcel must be scanned; a previous parcel’s label cannot substitute.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Provide colored paper parcels and two labeled bins before the digital run. Read labels aloud; color is also written in text. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Compare four explicit scan-and-sort pairs with the conditional routine. Explain why the conditional is reusable for a changed color order.

Independent exit check

Prompt: Can both sorting branches run for the same scanned parcel?

Teacher key: No. IF / ELSE selects one branch. A RED label chooses left; a BLUE label chooses right.

Connect to Scratch or physical robotics

In Scratch, represent color input with a variable or color sensor probe and use IF / ELSE. The simulator’s labels are authored task inputs, not camera recognition or AI.

Week 6: Alpine Rescue

50-minute lesson sequence

Model program and expected behavior

repeat 4 { move }; rescue; left
repeat 4 { move }; rescue; left
repeat 4 { move }; left
repeat 4 { move }; left; unload

The model route uses 16 moves, 4 left turns, 2 rescue actions, and 1 unload: 23 actions costing 21 battery units. One unit remains. Merely collecting both pods does not complete the rescue; unload requires both pods at base.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Give a perimeter map and a 22-unit token budget. Debit half-units with pairs of tokens or a number line. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Remove the final left turn before Unload: the rover can unload facing south, so 1.5 units remain. Compare a valid alternative without claiming it is globally optimal.

Independent exit check

Prompt: Why is reaching the second pod not the end of the rescue?

Teacher key: The rover must carry both pods back and unload at base. Plan battery for the return trip.

Connect to Scratch or physical robotics

A real robot’s energy use is not a fixed per-action token cost. Treat the battery as a simplified planning model and identify factors it omits.

Week 7: Orbital AI Trials

50-minute lesson sequence

Model program and expected behavior

Label four varied training scenes. Train.
For each unseen trial: predict -> review AI -> test its action.

A useful four-scene set is 100→Move, 010→Right, 001→Left, 000→Stop (bits = ahead/right/left clear). A learned tree then passes held-out 110 and 101 with Move and 011 with an open-side turn. Trial inputs cannot be used as training scenes. The independent rule checks the prediction; it never replaces it. A label change clears the current trial passes and requires retraining.

Misconceptions and a useful response

Differentiate the task

Support: Use path cards and one label at a time. Keep the explicit task rule visible. Allow spoken prediction and teacher-assisted selection. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Add the optional 111 training scene or deliberately mislabel one scene, then explain changes in tree paths and test results. Identify the tiny dataset as a limitation.

Independent exit check

Prompt: Does perfect agreement on training examples prove that a new prediction is right?

Teacher key: No. Test new inputs and independently check the action. Training-label agreement is not a probability of correctness.

Connect to Scratch or physical robotics

A Scratch IF / ELSE is an explicitly authored rule. This local supervised model learns a rule from labels. Neither this activity nor the task check is an online conversational assistant.

Week 8: Robot Maker Lab

50-minute lesson sequence

Model program and expected behavior

Attach a chassis, movement module, and sensor.
Test A: repeat 3 -> C4.
Test B: same start and module; repeat 2 -> C3.

All three module buttons change the robot assembly. Test A requires one tile per move, repeat 3, and a C4 prediction; Test B requires the same start and movement scale, repeat 2, and a C3 prediction. Wrong predictions or different parameters produce feedback and do not pass the calibration. The supplied model is not evidence of independently authoring a computational model.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Use counters and this grid as an alternative to interpreting the 3D view.

Misconceptions and a useful response

Differentiate the task

Support: Provide a five-cell strip and move a counter before each digital prediction. Let learners name modules orally. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Author a rover sprite and trail in Scratch, define move 40 as one tile, compare equivalent scripts, and test a changed repeat count. Save the authored project; this adds model-development evidence toward 4-6.CT.1.

Independent exit check

Prompt: Which inputs must stay fixed to compare repeat 3 with repeat 2 fairly?

Teacher key: Keep the same starting position, heading, and one-tile movement module. Change only repeat count.

Connect to Scratch or physical robotics

Scratch extension: create a sprite, start position and direction, movement scale, and repeat script. The browser module assembly alone is guided configuration; a learner-created Scratch model provides stronger authorship evidence.

Week 9: Mars Mission Designer

50-minute lesson sequence

Model program and expected behavior

Example for the starter route:
repeat 12 { if clear { move } else { left } }
Design changes need their own prediction and tests.

For an accessible starter design, add trail at C2 R4 and C3 R4, then place samples at C3 R5 and C5 R3. Test repeat counts 4, 8, and 12 with IF clear: Move / ELSE: Left. The final run reaches C5 R1 after eight moves and one left turn; goal checking ends the program early. Three distinct tests, two actual trail changes, at least two samples, and a successful loop-and-decision run are all required. Other maps may need other controllers.

2D teaching reference

Columns run west to east; rows run north to south. S = start, G = goal, M = mission item, X = barrier, P = route. Start heading: east. Week 9 shows example sample locations; learners place samples with the editor.

Misconceptions and a useful response

Differentiate the task

Support: Offer two suggested safe side tiles, but require the learner to place samples, predict tests, and explain the final revision. Use enlarged grids and dictated notes. Provide the written position/heading status and a 2D or tactile map when the 3D view is not useful.

Extend: Create a route where the starter left-turn controller fails. Design and test a different algorithm, then transfer the custom map to a learner-authored Scratch project.

Independent exit check

Prompt: What evidence supports your claim that the final mission design works?

Teacher key: A recorded map revision, three distinct tests, a successful final run, and an explanation of a limitation.

Connect to Scratch or physical robotics

The editable Mars map is a design artifact. Extend it in Scratch with an explicit goal, sensor probe, step limit, and saved project. Physical robotics would require different motion and sensor calibration.

Assess reasoning, not just a finished game

A short diagnostic before Lesson 1

• Ask the learner to follow three spoken or pictured instructions in order. Look for sequence, not reading speed.
• Show an east-facing arrow turning left. Ask the learner to move or rotate a counter to show what changes.
• Show one clap inside repeat 3. Ask for a demonstration or prediction.
• Ask what the word if changes in a familiar direction such as "if it rains, take an umbrella."
• Present an incorrect two-step route. Ask what information would help locate the mistake.

Use the diagnostic to select supports, not to rank learners or exclude them from programming. Record what they can explain with and without prompts.

Evidence collection across the unit

Four-level analytic rubric

Score each dimension separately from 1 to 4. Levels: 1 = needs substantial modeling; 2 = succeeds with prompts; 3 = explains and completes independently in this task; 4 = justifies and transfers to a meaningful new case. This is a teacher-created rubric, not a NYSED state scoring scale.

A suggested next step for a learner at level 2 is a smaller worked example followed by a new independent check. Do not turn a summed rubric score into a state proficiency designation. Pair work should include individual evidence.

Optional six-mission simulator: worked Mission 5 trace

C means column and R means row, both numbered 1-5. Energy is the total after that action. Turning stays on the same tile. Eight decisions produce six moves and two turns.

A useful final conference

• Point to one block and explain what happens before and after it.
• Choose one blocked case and explain the branch taken.
• Show one prediction that changed after testing.
• Explain one limitation of the controller or simulator.
• Name the next test you would run and what it would tell you.

Scratch and robotics transfer

A probe that samples only one point can miss narrow edges; sprite costumes, rotation centers, and color matching affect results. A single Scratch sprite cannot simply ask an arbitrary offset point for a color using the ordinary touching-color test. Test the probe implementation at corners and boundaries before use. Learners may first transfer only sequence, loops, and turns if a sensing model is not yet ready.

Physical robotics: adapt to the selected kit

• Choose the hardware and its supported programming extension before planning motor or sensor blocks. This guide deliberately avoids pretending one device-specific block set works for every kit.
• Have learners identify input, processor/controller, storage, and motor/display outputs. Map these to the four-part systems model.
• Calibrate forward travel and quarter-turns on a floor-level test area. Compare measured results with the simulator’s exact tile motion.
• For a distance sensor, collect readings at several distances and angles, then choose and justify a threshold. A range sensor does not, by itself, prove a floor edge is safe.
• Use low-speed, supervised tests with a clear stop procedure. Keep moving parts and cables clear, and follow the kit’s handling and setup instructions.

Optional data investigation

Use a teacher-approved spreadsheet to record decisions, forward moves, and turns across Missions 4-6. The supplied model programs produce 1/0, 6/2, and 10/4 forward-move/right-turn totals, respectively. Record the sensor-check count separately. Build a chart and explain what it shows. This provides additional work toward 4-6.CT.2 and 4-6.CT.3 when learners collect and organize digital data for an explicit question. See NYSED standards, grades 4-6 (p. 4).

K-12 progression: keep the task developmentally appropriate

Differentiation and classroom access

Collaboration and digital responsibility

Use respectful feedback: "I predicted..., I noticed..., could we test...?" Separate the role of typing from the role of thinking. Use first names or classroom identifiers on teacher-held records. Students do not need an account to use Rover Lab; any Scratch accounts, online sharing, recordings, or third-party tools should follow the school’s established procedures. The certificate name field is optional and should not be treated as a student information system.

Cross-curricular opportunities

Connect quarter-turns, counting, repeated addition, and route diagrams to mathematics; use claim-evidence explanations in speaking and writing; and connect sensor observations with scientific modeling. These are suggested connections rather than an additional formal mathematics or ELA standards crosswalk.

AI teaching notes: examples, predictions, and evidence

Four teaching methods used in the course

Explain the mechanism accurately

• Inputs are three Boolean map checks: ahead, right, and left. Clear indicates an available neighboring path tile; these are ideal map checks, not measured distances.
• Learners supply Move, Right, Left, or Stop labels. At least four different labeled training scenes and at least two action labels are required.
• Training fits a small decision tree by choosing feature splits that reduce label impurity. Each leaf predicts its most common action label; ties use a fixed ordering.
• The panel displays the learned decision path and supporting label count. Agreement is not a calibrated confidence or a probability of correctness.
• The independent task rule checks the output; it cannot replace a wrong prediction with a correct answer. Wrong labels can produce wrong model actions.
• The task is small: move if ahead is clear; otherwise turn toward an open side; stop if all three paths are blocked. Either turn is acceptable when both sides are open. This is not whole-route planning.
• Training and prediction run locally without an account, API key, voice recording, or external AI service. AI labels and trial passes reset on reload. Completion awards and worksheet responses can persist locally.

Training and held-out tests

The three trial combinations cannot be labeled in training. Passing these trials is evidence about this small test set, not a general guarantee. Changing a label clears the current model and trial passes; learners retrain and repeat the trials.

A useful counterexample

Label front-clear 100 as Right instead of Move, while labeling 010 Right, 001 Left, and 000 Stop. Train and test the held-out front-clear scenes. The model can suggest a turn that fails the separate task check. Correct the justified label, retrain, and compare the same tests. Adding arbitrary data does not guarantee improvement.

Prompts for discussion

• Before training: Which information is an input, and which is the action label? Why does the label fit?
• Before review: What do you predict for this new combination?
• After review: Which learned conditions led to this output? What does the label count mean?
• After testing: Did the learned action meet the task rule? What is your evidence?
• After revision: What changed in the training set, and what happened on the same test?
• Transfer: How is this learned rule different from writing a Scratch IF / ELSE yourself?

AI test record

Exit answer: training performance does not prove new-case accuracy. Test held-out inputs, compare against the task goal, and identify the tiny dataset and limited action task as limitations.

Supplemental sandbox

The optional older six-mission simulator retains a separate AI Rover Lab with freely chosen inputs and reviewed block suggestions. Its controls and saved examples differ. Use Orbital AI Trials for the three held-out 3D challenges and the Week 7 assessment.

Optional practice: six-mission simulator maps

These maps belong to the optional older Rover simulator, not the nine weekly games. The weekly lesson plans contain their own 2D references. Columns run west to east; rows run north to south. S = start, G = goal, E = energy, X = crate, P = permitted trail, and a dot = no permitted trail. GE means a crystal is on the goal; PX is a trail tile occupied by a crate. All starts face east. The numbered grid is a teaching aid.

Vocabulary, answer keys, and official references

Exit-card answer key

1. Pick up at C3 R3. Moving across the parcel does not collect it automatically.

2. It must also face north and have visited both beacons in order. Heading and position are different.

3. Eight actions: four moves and four processing actions. A shorter written program can perform the same work.

4. The scan is blocked, so IF clear skips its move. Confirm stop records the decision at the safe tile.

5. No. IF / ELSE selects one branch. A RED label chooses left; a BLUE label chooses right.

6. The rover must carry both pods back and unload at base. Plan battery for the return trip.

7. No. Test new inputs and independently check the action. Training-label agreement is not a probability of correctness.

8. Keep the same starting position, heading, and one-tile movement module. Change only repeat count.

9. A recorded map revision, three distinct tests, a successful final run, and an explanation of a limitation.

Official source documents

Standards documents were retrieved and the cited codes checked on October 18, 2026. Computational Thinking is on page 4 of each cited grade-band PDF; elementary Networks & System Design is on page 5 and Digital Literacy on page 7. Codes and brief summaries link to the original documents. Activities, timings, worksheets, and the rubric are original teacher-planning materials.

• NYSED: Computer Science and Digital Fluency Guidance & Resources
• NYSED: Computer Science & Digital Literacy FAQ
• NYSED standards, grades K-1
• NYSED standards, grades 2-3
• NYSED standards, grades 4-6
• NYSED standards, grades 7-8
• NYSED standards, grades 9-12
• Scratch: official ideas and tutorials
• Scratch: getting started guide

Prepared for New York Tutoring Center. Designed by Huanjiao Dong. Curriculum Created by Huanjiao Dong. Educator edition 4.0. This unit uses the New York State standards framework with a focused elementary scope and an extension roadmap across K-12.