Home   Ag Industry News

Killing weeds with robotics

Killing weeds with robotics

A look at experimentation with autonomous weeding equipment

Andrew Joseph

In the world of autonomous weeding equipment, it’s not about acres per day, but rather the number of weeds per day.

That’s the sentiment arrived at by Chuck Baresich, General Manager of the Haggerty Creek Ltd. and now Haggerty Ag Robotics Company, as he spoke at the Virtual Precision Agriculture Conference & Ag Technology Showcase held November 16-18, 2021. 

Baresich and his brother Justin have farmed in the Bothwell, southwestern Ontario area for 20 years in a long-term no-till zone system. From the beginning, they have used various forms of precision agriculture technology to farm. 

As part of an autonomous 2021 working group involving their Ag Robotics biz, and representatives from OMAFRA (Ontario Ministry of Agriculture, Food and Rural Affairs), the HMGA (Holland Marsh Growers Association) and others, the group included equipment manufacturers to its weekly meetings as the 2021 season progressed.  

The Autonomous Working Group 2021 then undertook its own research project. It created a plan to test autonomous vehicles in real-life crop production in Ontario, with an initial focus on autonomous weeding. 

Initially the group explored the following equipment:  

•    the GOAT from Nexus Robotics of Brossard, Quebec; 
•    the RoamIO HCT from Korechi Innovations Inc. of Oshawa, Ontario; and 
•    the OmniPower and OmniDrive systems from Raven Applied Technology (Canada) of Emerald Park, Saskatchewan.

As highlighted by Baresich during the presentation, other equipment was also added, including the Escalquens, from France-based Naio Technologies Oz and Dino robotic systems—distributed in North America by GMABE of St-Liguori, Quebec. 

Using robotics, laser visual sensors and artificial intelligence, these autonomous weeders were put through the paces at the Baresich farm. 

The Realities of Oz (and Dino)

Baresich applied the Naio autonomous weeders using only GPS for navigation across small, two-acre plots using multi-pass techniques. The Oz maximum speed is 1.8 km/h with an output of approximately 0.247 acres per hour. The Dino, for reference, could run twice as fast and had an output of 2.3 acres per hour. 

Where a crop was planted using RTK (real-time kinematic) position, the robots were used to mark A-B lines to navigate. Conversely, if no crop was planted, the robots’ RTK receiver was attached to the planter and an app was used to mark the A-B points for the robot to follow.

The fields tested contained Brussels sprouts, cabbage, carrots, celery, onions, peppers, tomatoes, sweet corn, cauliflower, asparagus and tree nurseries. 

Although weeding and tillage tools were provided by Naio—such as blades, hoes, harrows, tines, brushes and springs, Baresich said that while they sometimes performed extremely well, sometimes they didn’t. As such, new tools were fabricated that better-fit the soil types and conditions for Ontario.  

Baresich found that while some tools like the harrows performed very well, others could not break the soil up to the degree expected owing to soil type and compaction. He also found that heavy amounts of crop residue posed a problem for the robot systems.  

He noted, however, that robots are adjustable via tire alignment and tool configuration and can be adapted to many different row spacings.

Overall, Baresich reported that when the robots used smaller tines, each had difficulty breaking up soil crust in highly compacted soils. 

While larger blades left the soil overly clump for Baresich’s liking, he found that larger deep-set tines provided the best option on his land. 

For weeding itself—just as with standard weeding equipment, the autonomous weeders also had issues with weed roots clogging up the works. 

While using a deeper tool could resolve the issue, it was noted that it also caused the robots to veer off course. Baresich found, however, that along with extra weight and a set of dual tires, the issue was resolved. He mentioned that Naio has since developed an additional weight kit to counter that issue.

Overall, the Oz robots were straightforward to operate and understand, said Baresich, adding that the RTK GPS navigation system was “generally sufficient” in keeping the robots away from the crop while just attacking the weeds.
The RTK issues were minimal, he explained, using the Case IH RTK network via VRS. 

He said that the Oz worked fine with wet conditions—no issues—and could traverse most ruts. 

But how good was it in the actual weeding, you ask? Baresich said that the Dino and Oz performed very well. With regards to intercrop weeds, the robots were great controlling weeds between the rows, though sometimes multiple passes were required. 

He noted that weed capture was ideal when the weeds were small—which may imply that autonomous weeders can be applied early and often during the season to stay on top of weeds before growth becomes large and difficult. 

Crop growth patterns matter, said Baresich. The crops with a horizontal growth, such as tomatoes, have a limited weeding window, though that window could be extended if leaf guards were attached to the robot. Still, said Baresich, farms with vertical growth crops, or nurseries and berry farms could see great success with autonomous weeding technology. 

After all was said and done, Baresich said the Oz and Dino robots performed very well at the designated weed control assignment—especially since all involved were new to the technology. 

He summed up saying that it appeared as though the best process involved tackling small weeds with multiple passes. 

The working group will continue its meetings and experimentation in 2022, with plans to formalize the use of robotics systems with farmer cooperators. 

This article was included in the December 2021 Precision Agriculture Digital Digest — view it here.

Watch Chuck Baresich's presentation from the 2021 Virtual Precision Agriculture Conference & Ag Technology Showcase below.

Photo: Haggerty Ag Robotics Twitter

Trending Video

Protecting yourself from cyber threats

Video: Protecting yourself from cyber threats

In this video, learn how to detect suspicious online activity, create safe passwords and protect your devices and connections from cyber attacks.