I think what I would do here is simply have two global supply names, one for each supply e.g. +5V_A, and +5V_B. Then you can use the standard supply symbols.
OR
Have a master sheet with all the hierarchical sub sheets on it (power supply and driver boards) and wire things up directly using the hierarchical sheet inputs. Here is an example of this (only power net wired for clarity):
Master Sheet:
Individual Sheet:
Individual Sheet with two symbols used (Sheet 2):
PCB connection:
The same hierarchical label (3V) is used on each sheet, but on the master two separate regulators are used. One supplies sheet 1, and the other sheets 2 and 3. On sheet two another IC is also supplied with a separate 3V symbol - on the master sheet you can see two inputs are needed.
The hierarchical symbol does not appear to have automatic connection, so you either have to wire it up normally on that sheet, or add as many input of the same name to that sheet as separate symbols used.
You can on the PCB snapshot (the other net is a normal global ground symbol) everything is connected correctly.
In case this is relevant - if you want to split one supply into two nets, use a "jumper" component (e.g. 0Ω resistor) so the schematic doesn't complain, so then you can have e.g. main_supply, supply_1 and supply_2 all electrically connected, but split for PCB requirements (e.g. like you might have an analog and digital ground)
OR Possibly:
Make a power supply symbol, use the #
symbol in the reference designator (IIRC) which tells Kicad it's not a real component. Don't use a power flag on it though - this may work for a local power symbol if the quote below is correct (documentation is not the best though, and is outdated in some places so you need to be careful)
To quote from the link below:
A Power Symbol (VCC, V3P3, etc.) gives the net its name and is used on
each page to tie to the global power net. It is a special component
not listed in the BOM. A Power Flag (PWR_FLAG) symbol which gives the
net its global characteristics - connecting power nets between sheets.
There is info on creating power symbols at the bottom of this link.
Basically all non-trivial layouts will have mathematically guaranteed track crossings, if restricting yourself to a 2D plane. The trick is to not restrict yourself to a 2D plane ;)
Technique 1: Multi-layered PCB.
If track A needs to pass to the other side of track B, make it go through a via (an electrically conductive hole in the PCB) to the bottom PCB side, continue there for a bit, and come back up again through a second via on the other side of B. More sophisticated designs even utilize four, six, eight or more layers to make routing possible.
Technique 2: Jumpers.
Often you can get away with using parts (resistors/caps/ICs) as bridges. If the part is wide enough, you can make a trace go in between its pads and thus cross the path. If you do not have a part to route under, available where you need it, you can add so-called jumpers (think 0 ohm SMD resistors, or just pieces of wire for through-hole) to jump over existing traces. This solves the track crossing problem without utilizing extra PCB layers, and thus can save costs or time in PCB manufacturing.
Other tricks are for example connecting ICs not outward, but having traces go under the ICs body and continue from there. This can help disentangle the buses leaving thin-pitch ICs.
Best Answer
I believe you get the error at all the power in pins of your circuit: you are not providing them a power out connection since your power line is in series with the ferrite bead, which is passive/passive.
If you want to keep your filter symmetric, and it is symmetric, you can use the power flag. The power flag is basically a fake power output pin, where fake means it has no real counterpart. You place it on nets you know are powered to tell the DRC all is fine.
You can find the power flag component in the 'place power port' menu.