We created a robot called DTransleg that can walk on legs or roll on wheels using a simple, transformable design. Its upper leg doubles as a wheel rim, allowing smooth switching between walking and rolling. Tests with simulations and a real prototype show stable motion and easy transitions, paving the way for robots that adapt to different environments.

Farm robots currently struggle in scattered fields due to a lack of precise map data. To solve this, we created a system using satellite imagery and AI to automatically generate accurate field boundaries. Our tests showed 93 % accuracy across different regions. This technology serves as "digital eyes" for machinery, replacing slow manual inputs with automated data. It enables robots to navigate and harvest continuously in complex smallholder farms, unlocking the full potential of smart agriculture.