I feel like LNAV could be improved by being more predictive rather than reactive to the aircraft’s position relative to the next leg. From what I understand, the current transition generally starts around 1-2 NM before the waypoint, depending on the turn, and this can become problematic when the aircraft is fast or when the following waypoint is relatively close. For something like the A380, especially with sharper turns, this can leave the aircraft with too little room to make a proper turn and result in a much more aggressive correction than necessary.
Since LNAV already knows what track it is on, the next track, the aircraft’s groundspeed and the distance to the waypoint, it should be able to calculate the transition before reaching the waypoint. Instead of waiting until the waypoint area and then trying to turn onto the next leg, it could calculate the appropriate turn radius and anticipation point, create a temporary curved trajectory between the two tracks and follow that trajectory. The trajectory should update as the aircraft’s groundspeed and track change, so a faster aircraft naturally starts the turn earlier and uses a larger radius. Wind should also be taken into account, but I don’t think this needs live wind updating. A wind snapshot for the aircraft’s current cruise altitude would already be useful for estimating the expected groundspeed and drift and making the trajectory more accurate.
The bank should be progressive too, rather than immediately jumping to a large angle. It could smoothly start the turn, increase the bank only as necessary to follow the calculated trajectory, and then smoothly reduce the bank while capturing the next leg. There should also be aircraft specific bank and load factor limits, so the trajectory is based on what that particular aircraft is realistically able to do.
This is particularly relevant for the A380. In IF, I’ve noticed the A380 can initially reach around 1.3 g during some turns and then settle around 1.2 g. I wouldn’t use those numbers as a direct representation of what the real A380 normally experiences, but they do show how aggressive the current turn guidance can become. A wider and earlier turn should reduce that initial load instead of making the aircraft pull harder simply because the turn was started too late. For the A380, I would expect LNAV to favour earlier anticipation and a larger radius, with the bank remaining relatively shallow during normal managed turns. The A350 could receive the same general treatment with its own aircraft specific tuning, even though the A350 900 has not received the same level of rework yet. If the aircraft is already too close to the waypoint for the ideal trajectory, LNAV should still prioritise capturing the next leg, tightening the trajectory and increasing the bank only within the aircraft’s appropriate limits rather than missing the turn or oscillating around the route.
Course and radial intercepts
The same trajectory logic could also be used for course and radial intercepts involving VORs or NDBs. If the next part of the route requires intercepting a particular course or radial, the navigation system could use the aircraft’s current position and track, the navaid and required course or radial, together with the estimated wind and groundspeed, to work out where the intercept should happen and generate the transition path automatically. Instead of effectively flying a heading towards it, waiting for the radial or course and then correcting onto it, the system could calculate the intercept, create the transition and have LNAV follow that path until the aircraft is established on the required course or radial. Once established, it would continue tracking that course normally. This doesn’t mean turning everything into an RF leg. The idea is simply to use the available course and radial information to create a sensible transition between the aircraft’s current path and the required path. This could work across the whole fleet, not just the Airbus aircraft.
I don’t think this needs to become an extremely complicated FMGS implementation for a phone simulator. The main change could be a relatively simple trajectory generator between the existing route information and LNAV guidance. It looks ahead, determines the next required path, considers the aircraft’s current groundspeed and a wind snapshot at the current altitude, calculates the transition, creates the temporary trajectory and has LNAV follow it. The aircraft specific performance, bank and load factor characteristics can then be used to tune how that trajectory behaves.
I’m mainly suggesting this because I think the A380 and A350 need a little more Airbus specific treatment in their LNAV behaviour to make them actually feel like Airbus aircraft, while the underlying trajectory and course or radial intercept logic could still be used across the entire simulator.
Basically, instead of LNAV waiting for the waypoint and then chasing the next leg, let it calculate the transition beforehand, account for the expected wind, and fly through the turn.