Innovations


Recent Innovations Stories

Tomatoes grow in clusters, with stems and leaves that confuse harvest robots. Osaka Metropolitan University’s Takuya Fujinaga developed a system that estimates how likely a successful pick will be, then chooses the best approach direction. In tests, the model helped robots reach an 81% success rate, with many wins coming after the robot switched from a failed front approach to a side approach. The work points to a future where robots handle easy picks and humans tackle the toughest fruit.
New research shows robots can predict harvest success and change approach angles, reaching an 81% tomato-picking success rate.
Joshua Shavitupda