Overview
This early-stage project asks whether connectivity, local feedback and descending/ascending organization extracted from the MaleCNS v1.0 connectome can improve disturbance recovery and transfer between multirotor configurations.
Approach
The planned controller uses a shared recurrent core with perception and actuator adapters specific to each vehicle. The initial scope is multirotors with the same motor count and held-out mass, inertia and layouts. Planned comparisons use matched modular networks and rewired graphs under the same training budget.
Results & Current Status
Official MaleCNS v1.0 data have been downloaded and verified. Circuit selection, controller design and simulator selection remain research tasks; training has not started. Transfer across vehicle morphologies is a research objective, not a demonstrated result.
