The GUSS robot is a revolutionary innovation in agriculture, developed in California to automate spraying tasks in orchards and vineyards. This advanced system uses autonomous navigation and obstacle detection technologies to optimize the application of crop protection products.
In 2021, John Deere invested in GUSS, strengthening its expertise in precision agriculture and integrating cutting-edge autonomous technologies. Through this alliance, GUSS benefits from John Deere’s support network and advanced technologies, facilitating its adoption by farmers worldwide. Today, Future Farming visited a farmer who has been working with these technologies for over 3 years to gather his experience with these machines.
The GUSS machines arrived in 2020, almost four years ago. Since then, Ben Sill (Kern County, California, United States) has been using them to cope with rising labor costs and the increasing difficulty of finding available workers for this type of task. He manages the operation of the robots himself. The criteria that led him to choose GUSS include labor reduction, spraying efficiency and quality, as well as potential chemical savings, seen as the main benefits from a farmer’s point of view.
Since 2020, Ben has sprayed between 5 and 7 times a year on nearby fields, representing a total of 800 acres (320 ha). He keeps the robots close to the warehouse to avoid logistical problems – the 800 acres closest to the warehouse where the equipment is stored. In all, he grows 1,900 acres (640 ha) of almond trees. GUSS machines are very robust and powerful enough for spraying tasks, demonstrating reliable and durable performance in the field. Ben has always been impressed by the spray quality of these machines. Throughout the day, all he must do is “enter the spraying characteristics indicated on the products, and I know that the machine will apply exactly what is required, no less, no more.”
Ben manages the GUSS robots alone from his pick-up truck. He told us that it took him around four months and two to three spray cycles to fully understand and adapt to unpredictable situations and common problems. Operators have three very easy-to-use interfaces: an app on their phone, a remote control to move the machine remotely for logistics, and a portal for scheduling and tracking missions.
The operator is at the heart of the robots’ autonomy, taking care of the mapping of his plots. He must make sure he doesn’t forget anything and anticipates everything. The two main problems Ben encounters during these operations are the too-frequent safety-related stops, particularly the lidar, which mistakes branches and grass for living obstacles. Unfortunately, Ben practices regenerative agriculture, and his inter-row is sown with cover crops that can grow to over a meter in height. As a result, he must contend with numerous safety stops, as well as interruptions due to a lack of connection, which degrades the user experience.
Fleet operation demands a great deal of attention from the operator, as there is no real interactivity between the machines. Missions are programmed so that each robot oversees around ten rows, to avoid them crossing each other. The risk of bringing them too close together is that they will detect each other via lidar and jam. The operators then must refill over longer distances, and Ben must keep a close eye on the machines to deal with unforeseen events. Managing this fleet of robots is demanding and requires a lot of skills. This is why Ben finds it difficult to recruit suitably qualified staff to operate the fleet, which is necessary during peak periods, when he must spray day and night.
With the arrival of the robot fleet, Ben has changed his conventional operation with 6 conventional sprayers, six operators and 2 people in charge of refilling the sprayers. Today, he only needs two operators: one to manage the fleet of robots and another for chemical refills. With the 4 machines, Ben can spray between 15 and 20 acres per hour, an require around 400 hours of spraying per year.
Savings come mainly from reduced labor, machinery costs and spraying efficiency. With the fleet of robots, Ben guarantees precise application of products, independent of the operator, thus avoiding human error. Product and fuel savings are minimal, but noticeable at the ends of fields when robots evolve in an older orchard with trees missing. The connectivity of the GUSS system includes a subscription fee for the cellular connection.
Conventional organization:
That’s a total of $2,080,000 every 10 years
With the fleet of robots:
That’s a total of $1,790,000 every 10 years
Thanks to these 4 robots, Ben can reduce his labor needs and machinery costs, worth $290,000 versus a classic configuration over 10 years if you have around 400 hours of spraying per year. In conclusion, the pay-back period with these machines is enormously impacted by the number of people and machines you manage to save, but also by the number of hours you spend spraying each year.
The day after our visit to Silly Properties, we had the opportunity to take part in a contractor’s operation. On site, six GUSS machines were in operation in a 600-acre pistachio field, the operators expected to work between 30 and 40 hours to complete the entire plot.
Two additional trucks were present to prepare the chemical tanks, and two additional operators controlled the robots. The field was 1.5 miles long, and one round trip was enough to empty each machine. Each robot sprays 200 gallons of insecticides per acre, which means that they can cover covering 3 acres per load. The machines were operating at a speed of 2.5 miles per hour.
Text continues below picture
During our two visits, one point came up regularly: the triggering of the security system! In both cases, both operators were encountering too many stops due to the detection of an obstacle in the row. In most situations, these front lidar stops are false positives, covers crops too high or not mowed, a broken or downturned branch. In their opinion, these stops are too frequent, and Ben even told us that they deteriorate the user experience, as he must constantly check with cameras or visuals that there are no real obstacles in the machine’s path.
Given that no human is allowed in a plot during treatments and to avoid this fake positive problem, the contractor has chosen to reduce the sensitivity of the lidars. But to maintain a maximum level of safety, especially in the entry and exit of each row, two operators constantly anticipate the robots U-turns by getting there 1 minute early. The goal is to ensure that the robots do not encounter any obstacle in the headlands or in their next row. This method requires a minimum of two people for 6 machines, and much more concentration and organization to anticipate any eventuality, especially in the headlands, where other vehicles can be present.
These two testimonials raise the question of the relevance of lidars in an environment as unstable and unpredictable as an orchard. It seems that detecting something is not enough to be autonomous. The technology’s lack of discernment forces human beings to intervene on numerous occasions, reducing the autonomy of the machines.
Text continues below picture
The GUSS robot stands out for its integral control via a tablet, offering remote vision thanks to cameras and enabling the machine to be moved with a joystick if needed. A fleet of six robots, accompanied by three semi-trailers for transport and two trucks for chemicals, is operated by four people, two of them on quads or trucks. This configuration allows six drivers to be replaced by just two operators, increasing efficiency while reducing the doses applied between trees, especially when trees are missing.
Most of the times, the GUSS robot do not requires direct assistance in the rows, but, if necessary, the operator can control it remotely with a “PlayStation-type” joystick to avoid getting in touch with chemicals.
For two hours, we watched the uninterrupted comings and goings of machines, reminiscent of a well-orchestrated ballet. Nearby, the neighbor used two tractors for the same tasks, requiring frequent stops to refill at the station. The difference in productivity was impressive: not only were there four machines compared with six, but the trucks could anticipate and reload the robots as soon as they were empty, minimizing downtime thanks to effective anticipation and constant connection between the teams to optimize operations.
Subscribe to our newsletter to stay updated about all the need-to-know content in the agricultural sector, two times a week.