Darren Clemons
Well-Known Member
There's already been quite a bit of discussion on how the current "bucket" system of phase 1 of RTK works and what might be done to change it. Matt's mentioned of possibly being able to toggle to each individual bucket in the future and having some sort of scenario whereby the "worst" buckets were thrown out before it gets to 10 and kicks them all out.
As each day goes by and we work where we work, in mostly tough terrain and medium to heavy canopy we continue to see time and again, likely the "start" of a good solution throw out because of the 10 buckets being maxed out. Don't get me wrong, we're getting places we've never went before but with the LS programmers it always seems possible to get even better.
I realize putting the toggle ability on the buckets would put a lot more in the users hands and possibly take away from the basis of "letting it do its thing" but with enough experience on watching the epochs and learning how to read the data there's so many times the process could be faster. So many times we're following an existing survey and will have design points on every deed around our perimeter. Once adjusted, we'll have a point to stake to and depending on the survey expect to be within a few hundredths to a few tenths of the next point. If started in stake mode and buckets start to fill up, only one solution out of a possible 10 collected can/will be showing us data (the heaviest weighted one of course based on the current algorithms in the software), the other 9 "coordinates" if you will, are useless. Nothing can be viewed or used in them until their number of epochs becomes the largest in the batch. As we've most all figured out, a bucket with 3 epochs in 88 seconds is much more likely to be correct than one with 6 epochs in 6 seconds. It would be very helpful to be able to select each individual "bucket" to view what the stakeout solution would be for each. If any of them staked me right in on my design point, and if I had the ability, I'd kick it onto phase 2 of the verify process at that point. Sure, one or two epochs are a mile away from where I want to get before I'd ever accept and store a new coordinate but it would quickly get me towards what I have a pretty good hunch is right instead of spending 30 minutes filling up 1 epoch buckets over and over again (sometimes 30/40 before it goes to phase 2). By the way we use 120 seconds, 9 confindence and 20 consistency as our normal settings. I've toyed with lower confidence to get to phase 2 quicker, but it's a trade off of a few more that gets kicked out there and then starting over.
How about, if the toggling ability is ever implemented, having a simple setting to allow or not allow the user to toggle through the different solutions shown in the buckets? This way, if someone isn't comfortable with that process or if you send some of your crew out just to stake in points or collect boundary and you want to be sure what was done leave the toggle ability off.
Another possibility we've thought of on how to eliminate the ten buckets getting full and starting over is to just ignore the 1 epoch buckets. Don't show a bucket until it collects at least 2 (maybe even 3) epochs that match. In moderate to heavy canopy I'd bet 8 out of every 10 buckets is a 1 or 2 epoch count. Granted, in what I described above those can/could be useful, but as it is now without the ability to even use or see the "coordinate" that 1 epoch generates, they are not.
I'm not sure how, or even if, this could be done since the whole process of phase 1 is resetting after every single epoch. Could the LS somehow store the single epochs internally but not show it the bucket list until it gets a second or third epoch to match (an option of "minimum number of epochs per group to show in list")?
I'm sure there have been other things thought over and talked about, but the point is, there's a lot of data there that the LS goes through during phase 1 of verify and 9/10ths of every dumped phase 1 the user can't do, see or use any of that data.
As each day goes by and we work where we work, in mostly tough terrain and medium to heavy canopy we continue to see time and again, likely the "start" of a good solution throw out because of the 10 buckets being maxed out. Don't get me wrong, we're getting places we've never went before but with the LS programmers it always seems possible to get even better.
I realize putting the toggle ability on the buckets would put a lot more in the users hands and possibly take away from the basis of "letting it do its thing" but with enough experience on watching the epochs and learning how to read the data there's so many times the process could be faster. So many times we're following an existing survey and will have design points on every deed around our perimeter. Once adjusted, we'll have a point to stake to and depending on the survey expect to be within a few hundredths to a few tenths of the next point. If started in stake mode and buckets start to fill up, only one solution out of a possible 10 collected can/will be showing us data (the heaviest weighted one of course based on the current algorithms in the software), the other 9 "coordinates" if you will, are useless. Nothing can be viewed or used in them until their number of epochs becomes the largest in the batch. As we've most all figured out, a bucket with 3 epochs in 88 seconds is much more likely to be correct than one with 6 epochs in 6 seconds. It would be very helpful to be able to select each individual "bucket" to view what the stakeout solution would be for each. If any of them staked me right in on my design point, and if I had the ability, I'd kick it onto phase 2 of the verify process at that point. Sure, one or two epochs are a mile away from where I want to get before I'd ever accept and store a new coordinate but it would quickly get me towards what I have a pretty good hunch is right instead of spending 30 minutes filling up 1 epoch buckets over and over again (sometimes 30/40 before it goes to phase 2). By the way we use 120 seconds, 9 confindence and 20 consistency as our normal settings. I've toyed with lower confidence to get to phase 2 quicker, but it's a trade off of a few more that gets kicked out there and then starting over.
How about, if the toggling ability is ever implemented, having a simple setting to allow or not allow the user to toggle through the different solutions shown in the buckets? This way, if someone isn't comfortable with that process or if you send some of your crew out just to stake in points or collect boundary and you want to be sure what was done leave the toggle ability off.
Another possibility we've thought of on how to eliminate the ten buckets getting full and starting over is to just ignore the 1 epoch buckets. Don't show a bucket until it collects at least 2 (maybe even 3) epochs that match. In moderate to heavy canopy I'd bet 8 out of every 10 buckets is a 1 or 2 epoch count. Granted, in what I described above those can/could be useful, but as it is now without the ability to even use or see the "coordinate" that 1 epoch generates, they are not.
I'm not sure how, or even if, this could be done since the whole process of phase 1 is resetting after every single epoch. Could the LS somehow store the single epochs internally but not show it the bucket list until it gets a second or third epoch to match (an option of "minimum number of epochs per group to show in list")?
I'm sure there have been other things thought over and talked about, but the point is, there's a lot of data there that the LS goes through during phase 1 of verify and 9/10ths of every dumped phase 1 the user can't do, see or use any of that data.
