ANAIS: Advanced NAvigation Innovative System

[New Horizons]

Is anais gonna ever have something where you can just time warp to the most efficient transfer? Or like an indicator of where the most efficient transfer is, like say coloring the trajectory line.

Here's an example, like using existing colors to color your trajectory line based off of when the transfer would be most efficient. I'm probably stating things wrong lol. View attachment 118265

Amazing idea!

 

[Altaïr]

Is anais gonna ever have something where you can just time warp to the most efficient transfer? Or like an indicator of where the most efficient transfer is, like say coloring the trajectory line.

Here's an example, like using existing colors to color your trajectory line based off of when the transfer would be most efficient. I'm probably stating things wrong lol. View attachment 118265

The idea in itself is interesting, it could even be implemented, but at a very high cost in terms of performance. Currently, what ANAIS does is that it simulates transfers from the current position to the target at several possible arrival dates. If I was to implement this, it would need to simulate transfers for several arrival dates AND for several starting positions. If I make the calculation from let's say 20 different starting positions, the calculation time is multiplied by 20.
I take performance really in serious because this could be a real issue for such complex algorithms. Honestly I don't think the cost is worth it. I just managed to apply -40% on the calculation time, I don't see myself applying +2000% now

For informations, the calculations you suggest are similar to what is usually done to compute porkchop plots if you already heard about this, like this one:

In absolute this is possible to calculate, but not in real time. ANAIS does the calculations 30 times per second, this is not to be forgotten.

 

[Orion]

I'm having an issue where I can't type any values into the altitude option, everything just switches to zero.

 

[Altaïr]

I'm having an issue where I can't type any values into the altitude option, everything just switches to zero.

That part was developed quickly, so there could be some issues. Just to make sure, do you have a planet selected? Any value you enter is automatically reset to 0 when you type an altitude with no planet selected as target.

 

[Orion]

That part was developed quickly, so there could be some issues. Just to make sure, do you have a planet selected? Any value you enter is automatically reset to 0 when you type an altitude with no planet selected as target.

Oh, well that makes sense, that's probably the issue. Thanks!

 

[Altaïr]

Oh, well that makes sense, that's probably the issue. Thanks!

No problem, don't hesitate to tell me if it's unintuitive.
The reason why it's not possible to enter an altitude with no planet selected is because there's a sanity check: the altitude can't be above the SOI limit, but ANAIS needs to know the chosen planet to know the SOI radius. Also, I thought that if you entered a custom altitude then selected another planet the altitude you entered wouldn't be relevant anymore, so it's set by default to the time-warp limit (or the atmosphere limit if there's one).

 

[Darthan]

The idea in itself is interesting, it could even be implemented, but at a very high cost in terms of performance. Currently, what ANAIS does is that it simulates transfers from the current position to the target at several possible arrival dates. If I was to implement this, it would need to simulate transfers for several arrival dates AND for several starting positions. If I make the calculation from let's say 20 different starting positions, the calculation time is multiplied by 20.
I take performance really in serious because this could be a real issue for such complex algorithms. Honestly I don't think the cost is worth it. I just managed to apply -40% on the calculation time, I don't see myself applying +2000% now

You could do the calculation for alternate starting positions at a lower frequency. For example calculate one additional starting position 10 times per second (or even slower) storing the result in a list together with the time, and delete older entries as the list fills up. Also, if you did a pair of points at the same time , that are fairly close to each other, you could chose the next pair in the direction of the 'better' one of the two. It might be possible to converge on the 'best' position fairly quickly.

 

[Altaïr]

You could do the calculation for alternate starting positions at a lower frequency. For example calculate one additional starting position 10 times per second (or even slower) storing the result in a list together with the time, and delete older entries as the list fills up. Also, if you did a pair of points at the same time , that are fairly close to each other, you could chose the next pair in the direction of the 'better' one of the two. It might be possible to converge on the 'best' position fairly quickly.

Possible but complicated. In the sense "adds some complexity to something already complex". And it would probably be a problem in time-warp mode, if the ship orbits so fast that ANAIS doesn't have enough iterations available to locate the minimum over a period. However if it's just about calculating the optimal starting point it's not that complicated, I just need to simulate Hohmann transfers from several points and see which one leads to the target. This particular case is ok, but calculating many intermediate points would be very costly. Even if there's the possibility to spread the calculation over time, but at the price of increased complexity. I consider CPU time as a precious resource. Believe me, your CPU will thank me

 

[Darthan]

However if it's just about calculating the optimal starting point it's not that complicated, I just need to simulate Hohmann transfers from several points and see which one leads to the target. This particular case is ok, but calculating many intermediate points would be very costly.

I think that case is the one in demand. The builtin navigator shows the transfer window (or rather point), allows you to time warp to it (and shows the dv needed when you get there). That is often enough for the simple cases.