2.1 km · fails vehicle checks
Market access: Requires 18 cm clearance; vehicle profile has 15 cm.LOCAL TASKS. MEASURABLE ROBOTICS.
Robotics for Nigeria.
Start with evidence.
Test Nigerian English and Pidgin commands, explore local road scenarios, and build reproducible robotics experiments. Run the tools in your browser and export the evidence.
Take the controlsInteractive navigation lab
THE TEST ENVIRONMENT
14 × 10 GRIDA small hotel. A simple mission.
RobotPlanned routeObstacleX: 03 / Y: 08
YOU GIVE THE DIRECTIONS.Ready
Say it your way.
Give the robot a destination in English or Pidgin. Watch it find a clear route.
OR PICK A DESTINATION
Ready for your first command.
MISSION PROGRESS0%
DESTINATIONReception
GRID STEPS0 / 5
Command history00
- Your commands will appear here.
SIMULATION, BEFORE HARDWARE.
A deterministic grid simulator with predefined English and Pidgin commands. No physical robot or live AI model is connected.
04 / NIGERIA FIELD SIMULATORPut the map
Put the map
inside the mission.
A map-aware scenario layer for Nigerian places and conditions. Explore a route here, then carry its declared context into a Webots world when hardware and a physics model are available.
ACTIVE SCENE / LAGOS
LOCAL PREVIEWLagos Island loop
6.5244° N3.3792° ENIGERIA / 09JA
Dense mixed trafficMarket edges, junctions and pedestrian hand-offs.
SCENE INSPECTOR READY
Lagos Island loop
Market edges, junctions and pedestrian hand-offs.
PLACELagos
COORDINATES6.5244° N / 3.3792° E
PREVIEWAuthored region path
LAYERS2 visible
MAP LAYERS
A geometric path replay. No route planning or physics is applied.
WEBOTS BRIDGE / DECLARED STATECarry the scene into
Carry the scene into
a robot world.
Export an editable, inspectable Webots .wbt starter world with local geometry and source notes. It is a starting contract for a future robot controller, not a live connector or a physics simulation.
9JAROBOTICS / BRIDGESCENE V1
// Lagos Island loop / authored study
place = "Lagos, Nigeria"
source = "9jaRobotics local-context fixture"
layers = ["roads", "landmarks"]
webots = "editable world export ready" Webots runtime not connected · .wbt export ready
Scope: Stylised, authored geography for scenario design. The Webots export contains static geometry. Add a robot, controller and validated physical properties before conducting robot trials.
LANGUAGE → INTENT / 24 AUTHORED TASKSMeasure the instruction.
Measure the instruction.
Before the movement.
Compare predicted actions with explicit expectations in Nigerian English and Pidgin. This sample measures instruction interpretation; the navigation lab tests routes.
indoor-language-v0.1DECLARED: authored-fixture · unreviewed · CONSENT not-applicable
EXPLORE THE TASKS01 / 24
“Go to reception.”
SCENE CONTEXT
A fictional indoor map has four destinations: reception, room 01, store and charger. “Room one” means room 01. No other rooms are defined. The robot is stationary, awaiting one instruction. Interpret the full instruction; a correction replaces the earlier destination. This is a language annotation, not a physical execution result.
Expected interpretationNavigate → reception
COMPARE PREDICTIONSA small test.
A small test.
Every miss visible.
Run the existing rule parser, or import predictions from a model you evaluated separately. The rule baseline reads the instruction only and does not use scene context.
Files stay in this browser. Maximum 1 MB. Missing predictions count as misses. Template actions are placeholders to replace with your model’s output.
Ready. Run the baseline or import predictions to see exact matches and misses.
Sample provenance: 24 authored, unreviewed examples; 12 in Nigerian English and 12 in Pidgin. These are development fixtures, not collected human demonstrations. Both split labels are public. No model is called here, and these scores do not measure physical robot performance.
05 / NIGERIA ROUTE INTELLIGENCEA shortcut can
A shortcut can
cost more.
Compare a fastest assumed route with one that respects vehicle clearance, flood depth, surface quality, observation freshness and battery reserve.
TEST THE DECISIONSCENARIO / 08 SEP 2026
Clearance15 cm
Start / reserve20% / 15%
Human observations are the moat. 9jaTesters can collect road surface, flood and charger reports with contributor review attached.
DEPOT → COMMUNITY CLINIC
Conditional passVehicle-suitable route
4.1 km · arrival 18.6%
+2.0 km to avoid the riskier shortcut Delivery depot → Outer junction → Raised crossing → Community clinic3 segments
Outer access road2.3 min · 0.24 kWh
CautionRaised road2.1 min · 0.26 kWh
PassClinic access2.1 min · 0.19 kWh
PassWhy the fastest route is rejected
- Market access: Requires 18 cm clearance; vehicle profile has 15 cm.
- Low crossing: Requires 16 cm clearance; vehicle profile has 15 cm.
- Low crossing: Flood depth 18 cm exceeds the 6 cm test limit.
AUTHORED LAGOS-CONTEXT GRAPH / 8 SEGMENTS
These observations and coordinates are fictional fixtures. A suitable result is conditional on the inputs; it is not live navigation, EV telemetry, a safety guarantee or a real ETA. Real collection can be versioned through 9jaTesters and evaluated through 9jaBench.
01 / THE THESIS
Language. Perception. Action.
Test what connects them.
A command only makes sense in context. Use scene-grounded Nigerian English and Pidgin tasks to compare intended actions with model predictions, then test routes in a controlled grid. The current sample contains authored examples with declared provenance and review status.
LOCAL CONTEXT.REAL-WORLD AMBITION.
02 / A PRACTICAL PATH FORWARDSMALL STEPS. REAL PROGRESS.
From human context
to robot action.
Three kinds of evidence, each with a different role. Human video is a starting point; robot sensor and action data comes later.
01BUILDING NOW
Capture human tasks.
Use 9jaBots to annotate permissioned phone videos and describe the scene, instruction and intended action. Record provenance and permission status.
Explore the task studio02AVAILABLE BASELINES
Evaluate the system.
Run a rule baseline, import your model’s predictions and inspect task errors. Edit navigation scenarios and export reproducible grid results.
Run the task benchmark03FUTURE MILESTONE
Connect robot evidence.
With hardware access, collect synchronised robot observations and actions. Richer simulation and physical trials will test what the grid cannot.
How the evidence connects03 / THE COMPANY
Nigerian expertise.
Connected tools.
9jaRobotics is the robotics venture of Ranked Technologies Ltd, founded by author and builder Eruo Fredoline. Ranked.ng and 9jaTesters are established businesses serving paying customers. The lab connects local task knowledge with repeatable software experiments and a path to hardware research.
RESEARCH YOU CAN BUILD ON
Read. Download. Reproduce.
Keep the task, source and result together as your experiment develops.
GET STARTED
Run your first robotics experiment.
A practical guide to commands, obstacles and reproducible navigation checks.
DATASET / JSONNigerian English and Pidgin tasks.
24 authored examples with scene context and explicit expected actions.
COLLECTION GUIDEPlan a Nigerian mobility study.
100 assignment templates covering road scenes, objects, chargers and review.
TRY A TASK. KEEP THE EVIDENCE.Put a system
Put a system
to the test.
Start with the authored task pack and compare a baseline with your own predictions. Every result should be traceable to a task.
Run a task benchmarkLearn how to start without hardware