RL Robotics LabSearch ↗
Lesson / authored

Week 9 · Improve: Choose the quantity you can observe

Session 4 of 4 · Close the loop · Phase 3

Plan about 15 minutes for explanation, 30 minutes for practical work and 10–15 minutes for documentation. A longer build may continue into the next session: stop safely, commit the current state and record the next check. Desktop simulations count as software evidence; label them clearly and record physical validation separately.

Engineering challenge

Can measuring the result reject a disturbance that a timed command cannot notice? This session focuses on choose the quantity you can observe.

Before you start

The previous week’s recorded baseline and Week 8, Improve. For later sessions this week, retain the preceding session’s files and predictions.

Equipment: Desktop Python, editor, paper and ruler; for physical work, the configured 3pi+ 2040, clear floor mat and hardware checklist. Week 3 additionally uses the separate low-voltage LED circuit described in its procedure.

For any motion, verify the stop button, short time limit and clear floor area first. Keep the wheels raised for a new device program until its commands and stop behaviour are checked. A hazard or uncertain input is a reason to stop and document, not to force the trial to finish.

Theory and mathematics

Choose the quantity you can observe

Speed feedback and position feedback are different loops. Holding equal wheel speeds can reduce heading drift; reaching a distance target requires tracking accumulated travel as well. An encoder estimates wheel motion and cannot directly observe ground slip. Independent ruler/heading measurements are still needed. A controller can improve its measured variable while worsening an unmeasured one, such as jitter or stop accuracy. Include a command cap, total timeout and sensor-validity stop before any floor test. More feedback does not automatically mean more reliable navigation.

Worked example — illustrative values

If k = 0.0001 m/count, ΔN = 100 and Δt = 0.1 s, measured speed is 0.10 m/s. A reference of 0.12 m/s gives error +0.02 m/s. With gain 1000 command units per m/s, the proportional correction is +20 command units.

Write the calculation in your notebook before running code. State which values you measured, which you assumed and which the program calculates. A correct numerical calculation cannot rescue an incorrect physical assumption.

Run and explain the model

The following is desktop Python, not a ready-to-run motor program. Download this week’s example, save it in your student repository and run python3 code/w09.py from the repository root. The same small model is reused across the week so you can learn it, build with it, test it and revise it.

# Desktop Python teaching example. Numerical inputs are illustrative.
reference = 0.12
speed = 0.0
for step in range(40):
    disturbance = 0.7 if step >= 20 else 1.0
    error = reference - speed
    command = max(0.0, min(1.0, 0.4 + 2.0 * error))
    speed += 0.2 * (0.3 * command * disturbance - speed)
    print(step, round(speed, 4), round(error, 4), round(command, 4))

Compare the unchanged baseline and your proposed change on the same inputs. Keep the original files so another reader can reproduce the comparison. If an exception appears, read its final line, identify the input or assumption that caused it and make the smallest explained correction. Do not delete validation merely to obtain output.

Understanding the model and its limits

The command includes an illustrative feedforward value 0.4 plus proportional correction. The plant moves 20% of the way toward its new steady speed each step, representing lag. After step 20, multiplying the plant speed by 0.7 introduces a load-like disturbance. The error printed for each row uses the speed before that update, while the printed speed is after the update; label this timing if you plot both. Compare with a fixed-command run using the same plant and disturbance. This normalized command model is not Pololu’s ±6000 device-unit interface.

Practical instructions

  1. Add checks for positive dt, finite speed and maximum experiment time to your own implementation.
  2. Compare the response when a recorded sensor sample stops updating.
  3. Make stale/invalid input result in a stop, not reuse of an old drive indefinitely.
  4. Write which baseline weaknesses this feedback can address and which require other measurements.

Experiment

Inject a stale sample and a zero dt into a desktop controller wrapper. Verify that both trigger a deliberate rejection/stop rather than a division or uncontrolled command.

Before testing, record your prediction, changed factor, measured response, fixed conditions and stopping rule. Save every attempted run, including failures, with a condition and source version. If hardware is unavailable, use an explicitly labelled synthetic/replay dataset and list the physical question it cannot answer. Do not invent completed trials.

Deliverable

Invalid/stale-input behaviour checked and a clear distinction between wheel-speed and ground-distance feedback.

Save notebook/w09-s4.md, the relevant code revision, raw CSV or test-case records, and one labelled diagram/plot/table. Link the files relatively from your notebook. Use the entry template and report guide.

Completion criteria

A documented failed prediction can meet the learning criteria. A missing physical trial must remain marked untested; software success alone does not validate the robot.

Reading and video

Reflection and next step

Which assumption most affected your result? Point to one observation that supports your explanation and one alternative explanation the evidence has not ruled out. Write a specific next test with a changed factor and measurable outcome, then proceed through the week’s Learn → Build → Experiment → Improve cycle.

← Previous sessionCourse roadmapNext session →