How can an autonomous robotic platform perform reliable greenhouse navigation, ripeness-aware harvesting, and targeted disease spraying while operating under unstable rural network conditions and minimal human supervision?

Dense foliage, variable lighting, high humidity, and signal attenuation complicate both perception and communication reliability. Delayed intervention directly increases crop loss and treatment cost.

The system must maintain end-to-end autonomy by connecting vision diagnostics, precise actuation, and resilient networking into one closed-loop framework.

Can one edge platform reliably decide whether to harvest or spray, compute stable spatial targets for a 4-DOF arm, and still preserve command reliability when primary links degrade? The RP evaluates this through AI accuracy, intervention quality, and SDN-managed failover behavior.