Challenge 1
Rebuild all three followers side by side — Follow Line Basic, Follow Line Prop, Follow Line PID, each with a speed input — calibrate once, then run all three over the long leg at one shared speed and record time and swings.
The mission 4 min
Three followers, one leg, real numbers — and an honest answer about which to use where.
You have built three line followers across this course: the two-state one in lesson 5, the proportional one in lesson 12, and the PID one yesterday. Today they race, and you decide which legs of the farm route deserve which.
The answer is not automatically PID. It costs tuning time, it has more ways to go wrong, and on a short hop between neighbouring stations it may not finish settling before the leg is over. Today finds out where it is genuinely worth it — which is a different question from which is cleverest.
Two legs, deliberately different 5 min
The marked fast leg is not enough today. You need a contrast, and lesson 25 gave you one:
- The long leg — your marked fast leg, with its curve and its straight. Probably a diagonal across the farm.
- A short hop — two stations near each other. The market barn at A to the vegetable stall at B along the top works well, or anything into the greenhouses at C.
Both get all three followers. The contrast between the two results is the whole lesson, and a team that only tests the long leg will draw the wrong conclusion from it.
The technique — a fair test, twice 12 min
There are two honest ways to compare followers and they answer different questions. Run both, because either one on its own misleads.
Comparison A — same speed for all three. Pick one speed all three can manage and run them at it. This measures control quality: how close to the line, how much weaving, how quickly settled. It puts them on level terms.
Comparison B — each at its own highest reliable speed. Take each follower up in speed until it stops completing the leg every time, then step back one. Run it there. This measures what it buys you in a round.
B is where PID actually earns its place. Its real advantage is usually not that it follows better at a given speed — it is that it stays usable at a speed the two-state follower cannot survive at all. That extra speed is seconds on the clock, and the clock is what you are competing against.
One thing must be identical across every run today: the mid value from your colour calibration. All three followers use it. Re-calibrate once, at the start, and do not touch it again — a mid value that shifts halfway through the lesson invalidates everything measured before it.
Battery matters more today than usual too, because you are comparing across a whole lesson rather than within ten minutes. If it drops noticeably, re-run the first follower at the end to see whether the numbers have moved.
Build it — three blocks, side by side 10 min
You need all three followers available at once. If earlier ones were edited over the top of, rebuild them — the two-state follower takes five minutes and you have written it before.
Name them so nobody has to guess: Follow Line Basic, Follow Line Prop, Follow Line PID.
Give each a speed input, so comparison B is a number change and not a rebuild.
Keep all three afterwards. Do not delete the old ones once PID exists. A program that uses cheap control on cheap legs and PID where it pays is a better program than one that runs PID everywhere, and after today you will have the evidence for which is which.
Program it — the results table 12 min
Twelve runs at minimum: three followers, two legs, two comparisons. Twice each if there is time, because one run of anything on a mat is an anecdote.
Draw the table before you start. Rows are follower and leg; columns are:
- Comparison A — leg time at the shared speed, and swings.
- Comparison B — the highest reliable speed, and the leg time at it.
- Completed? — did the robot finish the leg on the line every time. A follower that is fast and occasionally leaves the mat has not won anything.
Have one person run and one person record. A number remembered until the end of the lesson is a number that has quietly become the number somebody hoped for.
Tune it — read the table honestly 9 min
Now answer four questions from what you actually measured:
- On the long leg, how much time did PID save over proportional — in comparison A, and in comparison B? The two numbers are usually quite different, and B is the one that shows up in a round.
- On the short hop, did PID save anything at all? Often it saves a fraction of a second or nothing, because the leg ends before the difference matters.
- Did any follower fail to complete? That column beats every time column beside it.
- Where does the two-state follower still make sense? It is the cheapest thing to keep working and has almost nothing to go wrong.
Expect a mixed answer, and trust it. PID on the long fast leg, proportional on medium legs, and quite possibly the basic follower on short hops. If your table says that, your program should say it too. Using the most advanced tool everywhere is not engineering.
Run it — assign a follower to every leg 6 min
Take the route card from lesson 25 and write, beside every single leg, which follower it will use and at what speed. Every leg gets an answer.
Then update the program to match, and run the whole route once. Time it, and compare against the walked estimate you wrote in lesson 25.
That total is the number lesson 31 will push down, and the mixed assignment you just made is the thing being pushed. Get it written before the milestone, not during it.
The journal 4 min
This is the entry that justifies the whole block:
- the full comparison table — three followers, two legs, both comparisons, plus the completed column
- the time PID saved on the long leg, and the time it saved on the short hop
- the follower assignment for every leg of the farm route, with speeds
- the whole-route time, against lesson 25’s walked estimate
- one sentence on when a simpler controller is the right answer — written from your numbers, not from what you expected
This table is worth showing a judge. Competition judging asks teams to justify their choices, and “we tested three controllers over two legs and here is what each was worth in seconds” is a far stronger answer than “we used PID”.