Robotic Sensor Finds the Hidden Harvest

The problem: Robotic harvesting in dense agricultural environments is limited because leaves and clutter block visual input for camera-guided robot arms. Existing sensor options are either easily damaged in the field or are impractically expensive for widespread use.

The solution: With support from the National Science Foundation (NSF) and the U.S. Department of Agriculture, researchers at Carnegie Mellon University’s Robotics Institute developed a data-driven system that uses sound to locate crops hidden from view.

• SonicBoom uses an array of contact microphones inside a protective pipe to detect and analyze audio vibrations upon touching an object. 
• This allows the system to accurately triangulate the location and determine the 3D shape of a crop, like an apple, simply by touching branches, even when it is hidden from view.

The impact: SonicBoom provides a more durable and cheaper alternative to conventional robotic sensors, overcoming a major hurdle for agricultural automation. By enabling robots to locate crops hidden among leaves, this innovation can be deployed for tasks like pruning vines or harvesting ripe apples, ultimately boosting farm efficiency and productivity.

Go deeper: New Robotic Agricultural Sensor Could Revolutionize Farming