Our planet’s human population is expected to exceed 9 billion by 2050. To keep pace with population growth, worldwide agricultural production will have to dramatically improve yields. One avenue to improving yields is through precise and persistent plant monitoring via mobile robots. These mobile robots can be deployed throughout farm fields, monitoring and tending to plants on an individual basis. They can live amongst the plants for entire crop cycles, operating with minimal human oversight or interaction. When combined with multi-agent control algorithms and data analytics, such systems could prove transformative to agricultural production.
A critical difficulty when deploying mobile robots in farm fields is locomotion. Wheeled robots require large wheels to avoid getting stuck in mud or tangled, but large wheels lead to large power requirements and risk trampling the plants themselves. Legged robots are not mature enough to operate in such environments for long periods reliably. Unmanned aerial vehicles only have flight times on the order of tens of minutes, so persistent monitoring and treatment over large fields is impractical.
A novel solution to the locomotion problem is the use of elevated wires. Such wires can be hung over crop rows, and can be lightweight and easy to install. We are currently developing brachiating, or swinging, robots that can traverse such wires. Our initial prototype uses an underhanded swinging motion to traverse along cables and between them. The ability to transfer between cables is imperative – it allows the robot to cover a full two-dimensional space such as an entire farm field. A payload of cameras and sensors, and eventually even water or fertilizer, can be carried by the robot. In addition to mechanical design, we are studying energy-efficient locomotion algorithms that allow very low-power swing maneuvers. These optimal swing trajectories, derived through direct collocation, will greatly reduce energy consumption and may enable such robots to be solar powered in the future.