Skip to content
gunkslab

Primers

The technologies gunkslab is built from, and the ones met along the way, one page each, in plain words. A primer is about the technology itself: what it is, how it works, a quick way to try it, and where to learn more. How gunkslab uses it is in the tutorials, guides and reference. Read one when a tutorial or a guide leans on something you haven’t met yet.

  1. MQTTA small publish/subscribe messaging protocol: every program talks through one broker. In three levels: high school, college, university
  2. ROS 2The libraries and tools many robotics labs and companies build on. In three levels: high school, college, university
  3. Robot communication and control stacksFrom wires to Wi-Fi to messaging (MQTT, ROS 2 and DDS, Zenoh, ZeroMQ, gRPC, MAVLink): what each is for, and when to pick which. In three levels: high school, college, university
  4. The ESP32A low-cost microcontroller with Wi-Fi built in: two cores, and pins for motors and sensors. In three levels: high school, college, university
  5. Time-of-flight sensorsDistance by timing light: lasers, photon counting, and what fools them. In three levels: high school, college, university
  6. BatteriesPower, part 1: cells, series and parallel (2S1P, 2S2P), capacity, sag and balancing. In three levels: high school, college, university
  7. Encoders and odometryCounting wheel turns into distance, and why the count drifts. In three levels: high school, college, university
  8. Feedback controlHolding a speed by measuring it and correcting: P, PI and PID, windup, and tuning by hand. In three levels: high school, college, university
  9. Coordinates and posesWhere a robot is and which way it faces: frames, transforms, wrapping angles, and how heading error grows with distance. In three levels: high school, college, university
  10. Testing softwareSmall programs that check a program: picking cases, failing first, test doubles, properties, mutation testing and flaky tests. In three levels: high school, college, university
  11. Measuring performanceLatency, throughput and percentiles, benchmarking and profiling, Amdahl's and Little's laws, queues and the tail. In three levels: high school, college, university
  12. Statistics for comparisonBetter, or just lucky: spread and standard error, confidence intervals, paired runs on the same seeds, p-values, power, and overfitting the test. In three levels: high school, college, university