Pendulous stability as it relates to drones appears to be a myth if the valuable opinion of Tom Stanton is taken into account at https://www.youtube.com/watch?v=OYHCP3-mpxk is to be believed. Conventional aerodynamics however have always regarded helicopters as being stable for precisely this reason, whether true or not, though the difference between the two types of aircraft is as much a reflection of the fact that a main-rotor has considerably more leverage I suspect. Tom Stanton's video suggests that drones with a decidedly low C of G are set to struggle... a view I've heard from various sources involved in real-world experimentation with drones.
Most of those involved in building drones, in fact, evolved from an interest in RC helis and in fact many of them prefer the latter as a flying machine given the limitations of drones at least in so far as flying people around is concerned. Nonetheless the fact that eVTOLs will succeed and displace or at least supplement the helicopter ~ to the extent they will far outnumber them ~ is inevitable in the way that electric motors will inevitably replace combustion engines in cars.
Before the expense of launching a working prototype (even at scale) therefore, it may be worth reviewing the options... one of which appears above. Prototypes are ideally an exercise in step-change where they start as the simplest and lightest embodiment and accrue the more practical accoutrements only later on. The automobile is as good an exemplar as any, in fact, having started out as an open-topped horseless carriage.
Accordingly the planform on the right provides the same leverage in flight, although it reduces the dimensions of the airframe by a third at the same time. It will tho' involve the centre-section in being reduced to an eight-inch square containing the mannekin instead of a twelve-inch, although I think this is workable in the first instance. At full scale it represents a four-pillared outline of sixteen inches, ample for supporting my own frame at the outset for instance.
Rather than adding motors to the drone up top, therefore, it would suggest re-visiting the previous solution I've implemented in past experiments to address an insufficient leverage around the centre of mass viz. replication of the same airframe around foot-level as well as above head height ~ in fact at the earliest stages of development this solution was only abandoned because it pitched the overall dimensions of a people-carrying drone outside the allowable constraints of the GoFly competition back when.
A temporary setback, but then nothing worthwhile in life ever comes easy.