12/03/1991

Greenhouse Robots Under Development

AUBURN, Ala.—Transplanting bedding plant seedlings is a monstrous task for most greenhouse and nursery operators. But help may be close at hand through the work of an Auburn University scientist who is breathing life into a greenhouse robot.

Larry Kutz, an agricultural engineer working with the Alabama Agricultural Experiment Station (AAES) at Auburn University, masterminded a creation that combines a robotic arm, specially developed fingers, and a brain to perform at least one of the industry's most unpleasant deeds, transplanting bedding plants such as pansies, impatiens and petunias.

"Many of the tasks performed in greenhouse production are repetitive and monotonous," said Kutz, who began this project while working on his Ph.D. at Purdue University. "The big advantage robots have over manual labor is that the robot can do these tasks continually. It doesn't require coffee breaks and it can work in the humid environment of the transplanting house more easily than people."

Bedding plant seedlings are usually transplanted into growing trays January through April, though certain plants are transplanted in the fall. This task requires a sizable labor force for a brief period of time and, according to Kutz, greenhouse operators often have difficulty hiring people for this temporary, tedious and low-paying job.

Kutz, who continued pursuing his robotic idea after joining Auburn's Department of Agricultural Engineering faculty in 1986, recognized the need for automation in the industry and figured a robotic arm could do the task.

"We first had to prove that a robotic arm had the capability to reach all the areas of the seedling flat and also get some idea of how long it would take the robot to transplant the seedlings," he said. "We had to see if it was even close to the time it would take to manually transplant these seedlings, because that's what greenhouse operators would be interested in."

An IBM 7535 robot was utilized for development and testing and proved effective. Kutz has since equipped it with fingers and a brain.

The fingers are actually grippers that lift the plant and root ball. Several prototypes were tried before grippers were found that were both strong and flexible enough for the task. A water nozzle also was added to the arm to help loosen the root ball from the grippers and settle the soil into the growing flat.

"We then had to implement some sort of intelligence in the robot, some sort of sensing device," Kutz continued. "We could program the robot to find a location, but the device could not know if there was actually a seedling in the area it was harvesting."

According to Kutz, a plant may not be produced in each cell of a seedling tray, yet producers need full growing flats to satisfy their customers. Human transplanters can be selective, but a robot, unless it can sense the presence of foliage, might simply transplant an empty plug of dirt.

Kutz chose a small infrared sensor that was available commercially and also was rugged enough to withstand greenhouse conditions. "It's very small, about one half inch square and one quarter inch high," said Kutz. "We can actually place several on the gripper."

The robot has been tested on a wide variety of bedding plants, including vinca, geraniums, impatiens, tomatoes, dahlias and petunias. Results indicate that the size and growth characteristics of plants affect the robot's transplanting success.

"We have found that there are species of plants that are simply better adapted to mechanical handling compared to other varieties," Kutz said. For example, the tall, flimsy stems of tomato seedlings are not as well suited to mechanical transplanting as smaller plants such as marigold, vinca and geranium seedlings that have abundant foliage.

"There are still some relatively minor problems associated with the sensing aspect of the robot," said Kutz. "The cycle time and speed of the robot is something we still have to work on, too."

Another goal is to make these robots multi-use. "By developing a robot that can be reprogrammed for other tasks, it could be more affordable for producers because it could be used for more than one species or growing arrangement, and perhaps for growing in flats as well as in pots. That's our goal, to make it as flexible as possible so it can be used for several tasks."

Kutz noted that robots are already common in manufacturing firms, but these designs are not suitable for greenhouse production because they are expensive and provide greater precision than is needed in the greenhouse environment. Kutz's garden-variety robot would be more economical and better suited for hothouse work.

Bridget Behe, a horticulturist who has assisted in Kutz's AAES project, said the robot could certainly benefit Alabama producers. She noted that greenhouse and nursery crop production are important agricultural enterprises in Alabama and bedding plants are a fast-growing sector of the industry.

While such a device may seem futuristic, Kutz noted that several companies already are developing greenhouse robots, some using his technology, so the hothouse robot may be lending its hand to the horticultural industry in the near future.

-30-

By:
Katie Smith