I've said it before and I'll say it again, you can't do a vehicle-installed repeater. Repeaters must not be capable of operating while moving, and cannot have a mobile operation mode. It's written in the Part 95 definition of a repeater station. They must operate at a fixed position. The specifics of what defines a fixed station is a question I'm trying to raise to the FCC's attention.
There's no spectrum entering or leaving GMRS. Consider frequency availability in the cities. Public safety is still building out on T-band here around Los Angeles because there is no spectrum available on 700 MHz, 800 MHz, or 450-470 MHz. Frequency reuse on those three segments is intense. Likewise, there's a lot of frequency reuse and congestion on GMRS, and coordination is necessary to protect against interference. In the suburbs out of the coverage areas of the cities, GMRS appears unused. Likewise, T-band and 800 MHz also appear unused. That doesn't mean there aren't any users. You just can't hear them, especially from ground-level. FRS users can exit receive range after a few thousand feet; low-level repeaters typical of private use can exit range in a few miles. Same goes for simplex or building-mounted repeaters in Part 90. The FCC won't deallocate 452.600 because they didn't hear anyone at their Denver field office in the last 25 minutes. For that same reason, they won't deallocate 462.600.
Since there's already a lot of Part 95 462/467 MHz equipment out there in a poorly regulated (relative to Part 90) fashion, getting FRS and GMRS users to stay off those bands would be impossible. Removing FRS or GMRS would also kill an industry of unlicensed two-way radio manufacturing (which is a bigger industry than you think), threaten emergency preparedness for what is easily hundreds of thousands of people, and provide very, very little benefit to the FCC. Trying to change the service to illegalize the operating modes of the existing radios is too monumental a task, and is one of the reasons why FRS gained the 8 repeater output channels. The pressures of the license-by-rule system on FRS and fixed channel set of GMRS would actually favor expansion of the band over contraction.
Setting up a bunch of repeaters blindly across all 8 channels just creates a bunch of interference for the other licensed users of the bands. Just because a repeater is open doesn't mean it isn't interference. This applies moreso during emergencies. Packing a bunch of users into GMRS in an attempt to lead the FCC to believe an already alive service is still alive doesn't make any sense. Setting up a bunch of repeaters in that close of proximity both in frequency and in physical space would also create intermodulation problems, further polluting spectrum.
If you want to use wide splits and/or tiny mobile repeaters with appreciable output power, use Amateur spectrum in either a wide split on 70cm or crossbanded to 2m/900. That spectrum has already been made available for hobbyist use. GMRS isn't supposed to be a tinkering band, type certification tries to ensure equipment already works when it reaches GMRS spectrum. Amateur is also free from frequency coordination concerns on temporary setups, restriction on mobile duplex operation, and linking concerns.
Cavity filtering only costs about $100, and considering there's a transmitter and receiver sharing an antenna, skimping on filtering isn't a good idea. Trying to implement a miniaturized filter would cause greater harm with receiver desense than benefit from running high transmit power, and with the filter order you need to get acceptable isolation even at 10 MHz, you're gonna spend more in component cost and tuning labor than you will buying a sixpack cavity filter. Single-channel repeaters don't cost much more than $400 to build ($100 cavity filtering, $75 transmit radio, $75 receive radio, $100 repeater controller and interface; power supply and antenna fill the balance). Performance isn't great, but it's certainly acceptable. Getting the cavities tuned is about knowing the right people or knowing enough theory to make cheap tools work (such as SDR + noise generator). Knowing a friend with the right tools is the right path to take, since they'll know more about the nuances of making a repeater work well. Building your first repeater alone isn't something I'd recommend, they aren't plug-and-play solutions and there's a reason this stuff costs money.