The working MVP, the custom PCB stack behind it, the platforms we studied, and the exact plan to get CropSCOUT into Uzbek greenhouses and vegetable farms — precision spraying and crop monitoring first.
A real, driving, sensing rover — steerable 4-wheel base, controllable front arm, soil & air sensing, and a wireless controller talking to it over MQTT. Built and presented at the ASU Fulton Innovation Showcase in front of 500+ people.
Every PCB in the MVP was designed and assembled by us — schematic, layout, fabrication, bring-up and debugging.
The board pictured is the motor-driver node: an ESP32-S3 driving 4× TLE9201SG H-bridges over SPI, one of the 7 PCBs in the rover's distributed architecture. Asadbek led the wheel-motor subsystem.
Next revision: the main controller moves to a Teensy 4.1 — more headroom, CAN bus, real-time control — and after that a Jetson Nano joins the stack for onboard vision and autonomy.
The concept design defined what the field robot has to do before a single part was ordered: front arm with temperature / humidity / pH probes, camera mast, swappable battery, land-marker stake dispenser, and a smartphone app showing soil & crop data in real time.
The current drivetrain runs over UART — the controller streams drive commands to the motor nodes, and the rover reports back. This short shows it running on the bench.
The next revision scales this up: more motors — per-wheel drive plus actuated attachments — moving to CAN bus so every motor node sits on one robust field-grade network.
Before designing our own field chassis, we studied how the best small ground platforms are built — drivetrain and suspension, CAN protocols, payload interfaces — and now know them inside out:
Full SolidWorks assembly of the field-scale CropSCOUT: chassis, drivetrain, and every attachment interface — modeled, reviewed, and frozen for fabrication.
The machine is complete and running by the end of the month. 3D-printed parts, aluminium and steel machining, fabrication and welding; motors and the Teensy 4.1 stack; the precision-spraying module and the full crop-monitoring sensor suite — camera, soil and climate sensors. Software lands with it: ROS integrated, LiDAR in for perception and navigation, and the autonomy stack working end-to-end.
Nothing left to build — from here we put it on real plots and run it. The first number we owe our farmers is the one they asked us for directly: how many sotix an hour it can actually cover. Every pass after that is measurement, failure, and iteration on ground we do not control.
The next module on the roadmap, once the sprayer is proven in the field: automated weeding on the same base — the job our respondents named as their single largest time cost.