Tutorials¶
The tutorial pages are the main content of this site. Each tutorial is intended to explain:
the robotics problem
the main algorithms used for that problem
where those algorithms are used
how those algorithms compare
what they cost in compute and memory
what to look for in the live simulator
Running the live embeds
The embedded simulator frames resolve their base URL at runtime:
local preview on
localhost/127.0.0.1:http://127.0.0.1:3000/hosted pages:
/sim/
You can override that explicitly with ?sim_base=... on the tutorial URL if needed.
Typical local workflow:
run
./build_web.sh --fastrun
./start_server.shrebuild this docs site with
./scripts/build_docs_site.sh
Tutorial structure¶
Each chapter is organized around four layers:
the robotics problem itself
the mathematical formulation and key assumptions
the computational tradeoffs such as time complexity, memory use, and runtime behavior
the observable effects in the interactive simulator
That structure matters because robotics learning often breaks down at the gap between the equation and the actual behavior. The simulator is there to close that gap.
This is deliberate. The goal is to keep the notes useful both as learning material and as a practical reference.
Linearization, feedback, optimal control, and reinforcement learning viewed through an inverted pendulum that is small enough to reason about by hand and rich enough to expose meaningful tradeoffs.
Bayesian state estimation, particle weighting, resampling, uncertainty propagation, and the practical limits of noisy sensing.
Graph search, heuristics, any-angle shortcuts, and sampling-based planning in a single environment where path quality and search effort can be compared side by side.
Joint pose-and-map estimation, covariance coupling, drift accumulation, loop closure intuition, and the difference between filter and graph viewpoints.
An end-to-end view of simulated robot control: observations, actions, policy execution, timing, portability, and what changes when an algorithm moves from a toy example to a more realistic runtime loop.
Suggested order¶
If you want a structured learning sequence:
start with control systems
continue with localization
move to path planning
then study SLAM
finish with the robot tutorial to connect the ideas to a richer runtime
That order roughly follows increasing model complexity and increasing coupling between estimation, planning, and control.
What these tutorials are not¶
They are not meant to be exhaustive research surveys, and they are not meant to mirror every internal UI option in the simulator. The goal is to explain the core algorithms clearly and compare them in a practical way.