This lab works on phones, but writing code is much easier on a laptop.
Localization & SLAM · Mathematical Foundations · M2-L4 · Intermediate
The robot sees seven blobs. It has eight landmarks on its map and a rough idea where it is. Which blob is which landmark, and which blobs are nothing at all?
You write the matcher, make it refuse to guess when two landmarks fit, and find the gate size that keeps the filter honest.
You have passed every stage. Now connect it back to the theory and move on:
Plain Python with math and, for matrices, slamkit (matmul, transpose, solve, inverse, identity). The data arrive as the dictionary D. There is no NumPy, random, json or copy in the browser runtime. Your code runs locally and never leaves this page.
Stage 1 is the standard gated nearest-neighbour matcher built on the Mahalanobis distance of M2-T7. Stage 2 adds the safeguard that matters in dense scenes: do not decide when the data cannot. Stage 3 runs a whole filter and shows how the gate size trades missed detections for wrong matches. The simulator knows which detection came from which landmark; your code only sees the numbers.
Educational simplification: point landmarks with known identities in the map, Gaussian sensor noise, uniform clutter, a pose-only filter and a greedy one-to-one assignment. Production systems also use appearance cues, joint compatibility (JCBB) or RANSAC, and track many hypotheses.
Get notified when new robotics labs are released. One email per new lab, nothing else.