
Engineer & validate.Line-Following Robot
Reproducible experiments, model pipelines and system design. Showcase build: an AI-assisted line-following study with a baseline and failure tests.
- 16 projects8 robotics, 4 AI, 4 integrated
- 20 to 80Sessions by plan
- Python + ArduinoMain language
- Supervised linkAI to hardware
What students do, and what they prove.
Line-Following Robot
An AI-assisted line-following study with a baseline and failure tests.
Learning check. Deliver a reproducible AI and robotics prototype with a baseline, failure tests and a model card.
Individual check. Every learner, on their own: “Defend a baseline comparison.”
Entry task
Use functions and datasets, and explain train/test separation and a robot stop state.
Robotics core
Line sensing, calibrated control, PWM, event logs, state machines and system integration.
AI layer
Supervised models, cross-validation, leakage checks, text features and reproducible evaluation.
Code & tools
Python and Arduino C/C++; optional SQL and JavaScript. Python + scikit-learn, Arduino IDE, Local datasets, API simulator.
Cross-subject link
Distributions, metrics, proportional control and experimental design.
Scaffolding
Template pipelines first. Advanced extensions only after demonstrated readiness.
How AI meets hardware here
Local inference and bounded serial control. Connected designs are simulated, not put online.
Sixteen projects, four blocks of twenty sessions.
Each block follows the same pattern: two builds, one AI project, then one integrated project. Open any project to see its five sessions.
- RRobotics
- AAI
- IIntegrated
- 01Ask & understand · Week 1Voltage, ground and safe motor power
- 02Build & design · Week 1Read individual line channels
- 03Implement · Week 2Record light/dark values and thresholds
- 04Test & improve · Week 2Test surface and ambient-light changes
- 05Explain & reflect · Week 3Explain a calibration table
Grade 11 UNO, line sensor, USB, test surfaces
Each sensor tested over light/dark surfaces; motors disconnected.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 06Ask & understand · Week 3Boolean decisions and motor directions
- 07Build & design · Week 4Build a basic line-state table
- 08Implement · Week 4Program a slow line-following loop
- 09Test & improve · Week 5Test missing line and motor stop
- 10Explain & reflect · Week 5Document baseline track performance
Grade 11 LFR kit, verified L298N module, approved pack
Three track runs and five lost-line tests.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 11Ask & understand · Week 6Features, labels and session-based splitting
- 12Build & design · Week 6Build a small line-sensor dataset
- 13Implement · Week 7Compare a baseline with a simple classifier
- 14Test & improve · Week 7Evaluate a separate capture session
- 15Explain & reflect · Week 8Explain leakage risks
Local Python/scikit-learn, recorded sensor CSV
Split by capture session; compare per-class results.
A model, an evaluation study or an AI-checking workflow.
- 16Ask & understand · Week 8Map predictions to bounded state names
- 17Build & design · Week 9Use a simulator before any motor link
- 18Implement · Week 9Compare predicted and rule-based states
- 19Test & improve · Week 10Require operator approval for physical runs
- 20Explain & reflect · Week 10Test stale data and unknown states
P02/P03, USB or simulator, approved stop control
Twenty recorded states plus invalid/stale cases.
A laptop model sends named commands only after bench tests, with a simulator as fallback.
- 21Ask & understand · Week 11PWM and load limits
- 22Build & design · Week 11Create a controlled travel experiment
- 23Implement · Week 12Record response at three settings
- 24Test & improve · Week 12Compare motor imbalance
- 25Explain & reflect · Week 13Select a conservative control setting
LFR chassis, ruler/timer, approved pack
Three settings, three runs each; no unguarded stall test.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 26Ask & understand · Week 13Error from a line-sensor array
- 27Build & design · Week 14Map error to bounded speed corrections
- 28Implement · Week 14Tune one gain cautiously
- 29Test & improve · Week 15Compare with the Boolean baseline
- 30Explain & reflect · Week 15Explain overshoot and loss of line
LFR kit, compatible multi-channel sensor output
Same track and start condition for both controllers.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 31Ask & understand · Week 16Train/validation/test roles
- 32Build & design · Week 16Build a preprocessing-plus-model pipeline
- 33Implement · Week 17Compare two simple models with fixed settings
- 34Test & improve · Week 17Check capture-session leakage
- 35Explain & reflect · Week 18Save configuration and results
Local Python/scikit-learn, sensor CSV
An untouched final test session; record random seed and versions.
A model, an evaluation study or an AI-checking workflow.
- 36Ask & understand · Week 18Define a fair comparison question
- 37Build & design · Week 19Run both decision methods in simulation
- 38Implement · Week 19Apply reviewed bounded serial outputs on a raised rig
- 39Test & improve · Week 20Measure error and response delay
- 40Explain & reflect · Week 20Decide whether ML adds value
LFR kit, laptop model, safe bridge or simulator
Same held-out route states and same safety limits for both methods.
A laptop model sends named commands only after bench tests, with a simulator as fallback.
- 41Ask & understand · Week 21Data schema, units and event timestamps
- 42Build & design · Week 21Log sensor states and commands
- 43Implement · Week 22Reject malformed messages
- 44Test & improve · Week 22Visualise a recorded run
- 45Explain & reflect · Week 23Explain why local serial is not internet IoT
UNO, USB, local Python, LFR kit
Fifty records including five malformed examples.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 46Ask & understand · Week 23Startup, running, lost-line and stopped states
- 47Build & design · Week 24Draw event and transition rules
- 48Implement · Week 24Implement timeout and manual isolation
- 49Test & improve · Week 25Test state priority and reset
- 50Explain & reflect · Week 25Publish an operator checklist
LFR kit, physical power switch, local console
All transitions plus five intentional faults.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 51Ask & understand · Week 26Define supported maintenance intents
- 52Build & design · Week 26Use synthetic short text examples
- 53Implement · Week 27Train a local text classifier
- 54Test & improve · Week 27Test unknown and contradictory requests
- 55Explain & reflect · Week 28Link results to source-backed checklists
Local Python, reviewed synthetic phrases, device notes
Twenty held-out cases; unsafe requests routed to a teacher.
A model, an evaluation study or an AI-checking workflow.
- 56Ask & understand · Week 28Design read-only tools and a command schema
- 57Build & design · Week 29Connect triage labels to approved troubleshooting cards
- 58Implement · Week 29Keep drive actions unavailable
- 59Test & improve · Week 30Test adversarial or irrelevant input
- 60Explain & reflect · Week 30Document permissions and refusals
P11 model, local manual, simulated tool API
Ten valid and five rejected requests; no hardware writes.
The AI advises a person. It has no control over the hardware.
- 61Ask & understand · Week 31Control experimental conditions
- 62Build & design · Week 31Build two clearly labelled track layouts
- 63Implement · Week 32Log repeated runs and route states
- 64Test & improve · Week 32Compare performance by condition
- 65Explain & reflect · Week 33Write a limitations statement
LFR kit, track material, logging computer
Three runs per layout with identical starting conditions.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 66Ask & understand · Week 33Device, gateway and dashboard roles
- 67Build & design · Week 34Specify JSON messages and status fields
- 68Implement · Week 34Simulate a connection using local files/loopback
- 69Test & improve · Week 35Test missing data and reconnect rules
- 70Explain & reflect · Week 35Explain what extra hardware real IoT needs
Computer, UNO optional, local JSON/CSV files
Five normal and five failure messages.
A physical circuit, sensor or mechanism, built and tested on the bench.
- 71Ask & understand · Week 36Define a task, trigger, resources and limits
- 72Build & design · Week 36Create a portable workflow specification
- 73Implement · Week 37Add examples, tests and version history
- 74Test & improve · Week 37Review optional tool scopes and dependencies
- 75Explain & reflect · Week 38Run a repeatability and failure test
Markdown/text, test files, local scripts; optional teacher Skills demo
Three valid and three invalid tasks with expected outcomes.
A model, an evaluation study or an AI-checking workflow.
- 76Ask & understand · Week 38Define a measurable robotics question
- 77Build & design · Week 39Integrate controller, model and audit log
- 78Implement · Week 39Preserve independent interlocks and human authority
- 79Test & improve · Week 40Run reproducible baseline and failure trials
- 80Explain & reflect · Week 40Defend results, limitations and responsible-use choices
LFR kit, local Python model, reviewed bridge or simulator
Three repeat runs plus unknown, stale-data, stop and reset tests.
A laptop model sends named commands only after bench tests, with a simulator as fallback.
Class 11 adds one more layer of judgement.
Reproducible system experiments
Pipelines, leakage and model comparison
Python and Arduino; optional SQL and JavaScript
A reproducible, evaluated workflow with failure tests.
Review dependencies, access scopes, logs and lifecycle.
A portable workflow specification and evaluation pack.
For school leadersBring Class 11
to your school.
Start with the four Discovery projects for Class 11 (20 sessions), review the evidence together, then extend to Builder or Innovation.
- Prospectus and class-wise plan
- Kit and readiness check
- Pilot one class first
- Evidence at every milestone
Let's plan your pilot.
Share a few details and we will send the right plan for your classes.