@FarmerDave convened session to discuss his tractor innovation and what he needs support with to integrate navigation. Folks in the group spent some time learning from farmer Dave about the existing and desired tractor tech (how it works, what’s involved, etc.) and shared thoughts/ideas/suggestions for what considerations might be involved in next steps for software integration.
Facilitator: Farmer Dave
(KR - formatting got a little wonky here, reach out if you’d like notes in a different format!)
Background/context:
• Farmer Dave won first place for CAFF innovation challenge for tractor innovation
• Don’t need to make my own rototiller, make a PTO that someone can use on a tractor
About the tractor:
• Current tractor is controlled by remote
• weighs 900 pounds, 7 inches wide
• Can connect to BCS implements
Target user - small farmers
• Currently using either hand equipment or 2 wheel tractor (BCS)
• someone trying to go bigger but can’t use compact tractor
• Farmers who want to switch to tractor for weed management
Value prop:
• For small farm allows more precision and control
• Discussion about benefits of going remote control to guided navigation
Affordability challenge:
• GPS needs to be inexpensive
○ FarmNG gps is 7500
○ John deer is 5k/year per tractor, not worth it, can’t justify expense
Farming Challenge:
• Weeds are biggest headache on farm, labor cost, or putting plastic down
• Easiest way to handle weeds - get them before they germinate, use spring tine rake
Dave’s ask/need:
• Wants tractor that can do supervised control (i.e. both remote control and guided navigation)
• We have the tractor, remote controller, antenna and we need to put inside that a controller which talks to the navigation system. Software needs to be integrated
• The nav system is a component inside the tractor and it runs ardu pilot
Existing tech to accomplish similar tasks:
• Low-tech option: use wire guide, mount remote on tractor….yolk turns it back
○ This works and still most cost effective, but is tedious to set up
○ Not easy driving a tractor down a straight line and weeding needs super straight line
• Competition - BCS
○ BCS doesn’t work b/c hard to use, not easy to make it go in straight lines
Technical convo:
• GPS is satellite, need a base station
○ Base station costs 3-4k
○ RTK corrects in real time, accuracy is higher
○ There is an open source base station
○ Plan is to map field with a robot, but start with survey company mapping it with waypoint coordination, gets plugged in, generates map
○ Question - what is the benefit of the GPS? More precise, allows for straight lines
Dave’s question to the technologists in the room: How difficult would this be to implement?
• Need to train data to recognize boundaries
○ This will require a lot of hours
○ Use highly reflective tape to make it obvious when tractor sees boundary
○ Needs to calculate distance and plan its way to stop at the end
• Use robot operating system
• Using GPS requires setting up lots of rules
○ One program runs to see what path it is taken (differs based on surface)
○ Range of allowed deviation
○ Every time it crosses that range it nudges back
• Outside versus inside tracking - need both
General group discussion/questions:
• Carbon robotics - use laser to weed
○ Cost barrier - 1.5 million dollars
• There is a cost to democratized tech - what cost are you comfortable with?
○ Tractor is 15k without autonomy
○ Want to put autonomy in for less than 5k and without or reasonable subscription fees
• How big is width of row?
○ Standard market garden farm is 48 inches center to center
○ 18 inch furrow, 30 inch bed top