<![CDATA[Anthracnose constrains chili production, and calendar spraying applies the same volume regardless of disease condition or canopy size. ANTRAC, an AI-enabled unmanned ground vehicle (UGV) developed at IPB University, detects the disease but has no actuation endpoint. This research developed and tested that endpoint, a spraying station that converts a fruit-level disease index into a metered spray command through four auditable steps: canopy-volume estimation, index-based tank selection, dose calculation bounded by safety rails, and pump-time derivation. An ESP32 controller serves a WebSocket endpoint, selects one of three isolated tank paths, and drives a ten-nozzle boom. Firmware output matched manual calculation in all three tested scenarios. For the 300 × 50 × 80 cm reference strip the lower rail raised the 75.2 mL raw dose to 180 mL, giving a 14.29 s command at 12.6 mL/s. Two hydraulic checks at the 560 mL command delivered 8.4–12.3% less liquid than commanded, locating the dominant error in the open-loop hydraulic path, not the arithmetic. The station held Wi-Fi association through 50 m on an obstructed path and accepted a spray command near −88 dBm. The tests establish station-side functional readiness, not closed-loop operation or agronomic efficacy.]]>
Copyrights © 2026
