Impact of system latency upon controllability ?

[Genghis the Engineer]

I've just been wading through a (non aviation) document on system latency that claims to be showing me some wonderful lessons about how critical system latency (input-response delay) is to controllability. No sh1t Sherlock!

However, when I went and had a rummage around the usual sources, I can't seem to find any good documentation to support what I believe I intuitively know either for aircraft, or the other obvious user of such principles - computer game design.

Can anybody point me at a good document on the topic?

G

[Fitter2]

When I passed through NASA Dryden (Edwards AFB) in '69, they were doing some serious research on that, related to deliberate crashing an unmanned aircraft in a controlled way (they didn't wan't an uncontrolled crash, for some reason). I can't find anything from then in my library, but the first reference on the link below may be of relevance.

https://tinyurl.com/y7deuqjz

[Genghis the Engineer]

Thanks Fitter, but I suspect it isn't. I can't see drone command signal latency having much relevance to the problem of control / FBW direct input response latency?

G

[aerolearner]

This looks interesting:
https://etd.ohiolink.edu/pg_10?0::NO:10:P10_ACCESSION_NUM:dayton1418340523
It presents a study carried out on a flight sim F-16 Flight Control System model, so I understand that the latency limits they found between a stable and unstable behaviour are related to the flight sim model they used and do not necessarily reflect what you would find on a real F-16. On the other side, I suppose the investigation method would be applicable to a real case as well.

The relevant paragraphs are 3.12 and 6.2.

Edit: The link to the .pdf file with the full text of the thesis is at the bottom of the webpage and not very conspicuous, so I enclose it below.
https://etd.ohiolink.edu/!etd.send_file?accession=dayton1418340523&disposition=attachment

[ICT_SLB]

Genghis,
Not my specialty as I more ensured that the overall AFCS met requirements & certification, but my recollection is that what you're asking was a fairly closely held outcome of programmes such as Bombardier's Active Control Technology (ACT) run in conjunction with Rockwell Collins' Flight Controls group. PM me if you need potential authors for a search.

A quick search on Latency on the SFTE website only brought up one paper but, again, my memory from around 10 years ago is that there were several such papers as smaller manufacturers looked at FBW. This is the link: https://www.sfte.org/component/content/article/101-services/papers/132-technical-paper?paper_id=1264

[ICT_SLB]

Genghis,
Just got this in an e-mail from the IET Digital Library.

Execution time analysis and optimisation techniques in the model-based development of a flight control software

This case study analyses the possibilities to improve the execution time of model-based developed software by applying optimisations during code generation and compilation. The present case study is performed on flight control software, for which safety aspects are accounted throughout the development. Therefore, a formally verified compiler is used for the optimisation during the compilation. The optimisation is evaluated by execution time measurements on the target and a static worst-case execution time analysis. Based on the results, recommendations for certain model patterns are given, which impact the worst-case execution time analysis.

http://digital-library.theiet.org/docserver/fulltext/iet-cps/2/2/IET-CPS.2016.0046.pdf?expires=1504063742&id=id&accname=guest&checksum=70D1D9D735F5AE1D9E3CD80308BCDE83

[212man]

My main exposure to latency has been in Full Flight Simulators. Some pertinent stuff here:
https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20050192646.pdf
http://reports.nlr.nl/xmlui/bitstream/handle/10921/713/TP-2002-470.pdf?sequence=1

[PDR1]

You may not find a specific reference because it's too basic. Any control engineering text looking at system dynamics using Laplace or similar makes this issue self-evident. Control systems implement negative feedback to achieve stability (a deviation triggers a control input in the opposite sense to the deviation). Latency is essentially phase-lag, and any phase lag in a negative-feedback system produces positive feedback (instability) resulting in either oscillation or divergence. That's the basic bit, but at deeper levels you can have nth-order phase lag (ie where the lag is in the velocity, acceleration or jerk response rather than the position response) which need to be understood mathematically for the correct fixes to be applied.

As a practical example one of the test pilot memoirs I've read talked about the Fairy Delta-II (I think) whose powered control system was rather experimental and suffered lags in the hydraulics until the oil warmed up (or something like that). So when each pilot flew it for the first time they would climb away with an apparently uncontrolable 1-2Hz pitch oscillation which would be rather alarming for the first five minutes, and then progressively disappeared.

PDR

[Genghis the Engineer]

Thanks guys.

I've bounced the document (a paper sent to me to review by a "leading" academic journal) a strongly worded recommendation that they go off, talk to all or any of test pilots, simulator engineers and/or computer games designers and re-write it in the context of a chunk of real world experience. 212's documents I gave a special mention to.

PPrune at its best!

G

[Phil_and_Sand]

Genghis,

Any delay in visual representation of actual (or preceived) motion greater than 100ms will make an aircraft more difficult to control. A (test) pilot will report the delay as degraded aircraft handling properties, but it is actually poor performance of the (simulator) visual system of the PFD generator. The corollary is than a delay of less than 100ms is likely to be acceptable. No reference available I'm afraid, just hard won experience.

Phil

[Genghis the Engineer]

Well indeed - and I think that most flight testers (along with simulator engineers, computer game designers...) know all of that fairly intuitively - as you and I do.

The big surprise for me here was the lack of formal experimental evidence of this. After all, I'd not have to look at all hard, for example, for data relating stick force gradients to controllability, or damping to risk of divergent oscillatory modes.

G