It is well documented that the LM could land on its own, that is by way of autopilot.
If only you knew what they are talking about!
As usual, it's not what you think it means. And as usual, you're quoting popular sources and failing to delve into the technical details that they leave out. I know them, and you don't. Hence you wallow in your own naive interpretations while the rest of the technically-competent world accepts Apollo as real. You hear the word "autopilot" and you get a set of ideas in your head, and without checking to see whether they're accurate ideas, you immediately start judging reality by them.
Nor will this be the first (or even the second, third, or fourth) time you've been entirely mistaken about LM guidance and control. Keep in mind that you come to this discussion with a huge degree of "fail" behind you on this topic.
PGNS employed three computer programs from Luminary -- P63, P64, and P66 -- to "autopilot" the landing. In fact, each of these required pilot input. They were not "auto" in the sense of unattended flight, but "auto" in the sense that many guidance commands were based on a numerical model rather than on pilot commands, thus freeing the pilot from tedium and error. "Autopilot" is never synonymous with "unmanned."
The most obvious error in your thinking is that each of these programs, at various times, asked the pilot to confirm its computations and intentions. The program is there for anyone to inspect. If there's no human pilot to press the PRO button, the computer program will not proceed. So much for your claim that the LM manufacturers wouldn't have known they were building what would be no more than an unmanned vehicle.
P63 first asks the pilot to confirm the ignition timing parameters. It then asks for final ignition confirmation. The pilot must then designate the landing site by manually setting Noun 69. During this phase, there is no automatic updating of the AGS to accommodate AGC-only information such as radar. The pilot must manually synchronize the AGS using hand-keyed inputs.
P64 also requires pilot confirmation to begin its work. P64 is completely capable of flying the LM from pitchover to terminal descent without any pilot input. But the pilot has two things the computer lacks: eyes. Only the pilot can determine whether the landing site to which the computer is sending them is a suitable one for his spacecraft. P64 will blissfully fly the ship into the side of a crater or on top of a pile of boulders. It's up to the pilot to tell P64 where to land. Noun 69 is only good to about 1,000 feet, and within that tolerance can be a lot of nastiness.
Finally we come to P66 (terminal descent), which is your Waterloo. It takes over at about 200 feet, again with pilot confirmation. Up until now the PGNS mode is AUTO, meaning that the DAP takes commands from the targeting program. In P64 the targeting program gets pilot input from the joystick and translates those into commands to adjust the spacecraft's attitude to land shorter, longer, left, or right from the previously designated landing point. To land shorter, for example, the DPS is throttled back slightly for a short period, allowing the spacecraft to dip lower.
But in P66 the PGNS mode switches to ATT HOLD and the DPS throttle mode switches to ROD HOLD. In ROD HOLD mode, the computer simply maintains a selected descent rate, with that rate determined entirely by the pilot. No numerical model applies here. P64 will leave the descent rate somewhere in the neighborhood of 12 fps, but there's no guarantee of it. The pilot is supposed to manage the descent rate from the 12 fps to the 2-3 fps at touchdown; 12 fps at touchdown will exceed the spacecraft structural limits.
Because the LM is expending fuel and thereby growing lighter, the throttle has to be closed slightly to maintain the constant descent rate. But it's still up to the pilot to determine what that rate should be. P66 is considered "autopilot" because it manages the throttle according to a high-level input from the pilot -- the ROD selector switch.
P66 provides no lateral control. That is, horizontal rates are entirely under the command of the pilot, who is responsible for nulling them out. That is the essence of ATT HOLD. The LM will maintain its orientation in space, regardless of whether it's a good orientation or not. Nominally P64 exits with the spacecraft moving slightly forward, but there is no guarantee of this. The pilot is meant to correct the final laterals. In P66 the pilot sets the orientation (and therefore lateral accelerations) with the joystick. He has a manual throttle command if he wants it, but the "autopilot" function of P66 is to manage that for him.
From your source: "It was the
LM pilot’s responsibility to notice when the contact light went on." (emphasis added) The LM designers deliberately made it the pilot's choice whether to act on the contact signal or not. There was no automatic cutoff. The pilot and the pilot alone made the decision to go for touchdown.
MIT wanted to enhance P66 to do a full automatic landing, but they were never allowed to. The original Luminary P66 was simply a throttle controller. The Luminary P66 as enhanced for the J-type missions provided an automatic mode for lateral stability, but still no automatic descent-rate model.
What is the whole controversy then between the pilots and the engineers? The pilot could enter P66 at any time, even before the 200-foot limit for P64. And most pilots did. At a certain point they decided that P64 had gotten them close enough. And in Armstrong's case, he realized that his need to avoid the boulder field was not something the LPD could address. Hence he manually commanded the computer to P66 at about 600 feet, leveled off, and flew horizontally in a way P64 could not do. While the other missions would probably have succeeded if they had let P64 take them all the way down to 200 feet, Apollo 11 would have crashed.
The context you're missing is that "autopilot"
under no circumstances was meant to control the spacecraft without pilot input below 200 feet or to provide for touchdown and engine commands. That was always the job of P66 and a highly experienced pilot who could be trusted to avoid obstacles and compensate for any altitude-sensing errors by visually judging the altitude and commanding the descent rate. Your source is talking about the premature termination of P64 in favor of P66. The engineers had wished P64 would be allowed to run to completion, placing the LM 200 feet above the exact landing site.
There really is no way you can handwave around the very evident fact that the LM was meant to require a human pilot, even if it nevertheless incorporated a high degree of automation.