2024 maverick excessive brake usage while using openpilot

### Describe the car bug

Ever since I used the comma3x and 4 I noticed some unpleasant side effects in regard to brake usage and fuel consumption. I am providing this information as feedback to the developers and not as criticism of their work.

1) fords autostop, which shuts of the engine when, for instance stopped at a light, does not engage as quickly if at all when the comma device is on or connected, Just verified this with the unit removed. When I stop at a red light, within 2 seconds the motor will shut off. With the c4 connected, this would rarely, if at all, happen. Mostly affected when in city traffic and no matter if the c4 is enabled or not, just being connected seems to have that effect.

2) excessive break wear. My OEM brake disks got warped and I had to replace them after only using open pilot for a few times. Put in aftermarket heavy duty disks and breaks, and noticed significant wear already. My observations show that OP seems to engage throttle AND breaks at the SAME time. Evidence for that is my fuel consumption indicator and the fact that my breaks make some noise when engaged. I noticed excessive braking in denser traffic and when going downhills on freeways. OP attempts to not only keep speed but also distance to a narrow window of margin and attempts this by engaging throttle and breaks frequently at the same time, thusly causing excessive wear. Now, this operation may be suitable for an electric or hybrid vehicle where regenerative braking is used before brakes are actually engaged but not for an ICE engine.
Which makes me wonder if I forgot some setting to tell OP that I am on an ICE vehicle.

The effect is very noticeable, I see fuel consumption on my dash board and feel deceleration from the break action along with the break noise. The vehicle seems to fight itself. As soon as OP is disengaged the vehicle is rolling relaxed in coasting mode. The effect is irrelevant from using fords longitudinal control versus the OP experimental longitudinal control and in all driving profiles (aggressive, relaxed, standard) while c4 is connected to the vehicle.
When I drove the vehicle without C4 connected, i notice that the fords longitudinal control does NOT show such “infighting” between acceleration and breaking and fuel consumption improves significantly.

3) items 1-2 have a direct result on fuel ecconmy. Without OP on level FWY I get around 32mpg , city about 26 with fords cruise control. With OP connected and controlling, this changes at best 24-25 FWY (without elevation changes) and 22-24 city.

4) infighting was also noticeable during my C3x time, I switched to sunny pilot at that time which resolved that wear to some degree which handled those situation apparently better and also allowed steering control without longitudinal control

5) earlier versions of OP longitudinal control were pretty bad in a way that made the car hesitated too much and sit at lower speed before deciding the return back to the dialed in speed. This seems to have Improved.

6) I would hope OP could engage more in coasting or pulse driving which reduces fuel consumption considerably. OP should also not insist of holding the speed or distance to close of their settings but allow a more relaxed variation AND avoid braking and acceleration at the same time.

### Provide a route where the issue occurs

a23987ab645edc87

### openpilot version

0.11.1

### Additional info

_No response_

Start by reproducing the reported behavior on the 2024 Maverick with openpilot 0.11.1 and route a23987ab645edc87, comparing the comma device connected and removed. Investigate the vehicle-control behavior across the reported longitudinal modes and driving profiles. Done means the excessive braking, apparent simultaneous throttle and braking, auto-stop interference, and fuel-consumption regression are explained and addressed.

Written by the indexing model from the issue text.