Building a rover for Nigerian dirt

The robotics track exists because the tools Ifenri needs do not exist at a price that makes sense for a two-hectare plot. But it starts with one narrow task, not a platform ambition.

The first robot is a small rover that follows a crop row using a camera and GNSS, logs what it sees, and stops safely. That is all. It has to complete twenty consecutive row passes in a test plot with no software-caused safety stops before it goes anywhere near a paying farm.

The constraints that actually shape the design are physical, not computational. Fine dust destroys bearings and clogs cooling. Heat degrades battery life and forces derating. Vibration loosens connectors that were tight in the workshop. Mud defeats wheels sized for a laboratory floor. And equipment left in a field overnight has a habit of disappearing.

So every part is chosen for whether a local technician can swap it in a day, and every enclosure is chosen for whether it survives a season. A robot that can be repaired in the cluster beats a robot with a better specification that has to be shipped abroad.

We publish the failure log. Writing down every failure and fix is how the second robot gets cheaper than the first.