Autonomous weeders for Christmas tree plantations - a feasibility study
10 Conclusions
- Current Christmas tree production occupies about 31000 ha and has an annual turnover of
500-600 mill. DKK. There are about 4100 producers, of whom the majority have less than 1
ha.
- Most Christmas trees are produced on farmland. Typically two year old trees are planted
on prepared soil. They are cut after 6 to 10 years using selective harvest. Planting
patterns are typically square grids with spacing in the range 90 cm to 125 cm.
- Present weed control measures are mainly chemical (about 70%) and the costs are about
20-25% of the total cultivation costs. Mechanical weeding methods are being developed but
are still relatively costly and can mostly only do inter row weeding, which has limited
effect and is causing nutrient leaching and soil erosion.
- The minimum requirement to weed control is found to be an area of about 40 cm radius
around each tree corresponding to about 40% of the total area.
- The requirements of stakeholders are that autonomous Christmas tree weeders should be
competitive to present methods and fulfil the minimum requirement of weeding. Stakeholders
also want to have several other work tasks, i.e. growth regulation, spot spraying, spatial
variable fertilizing, and data collection, done by the autonomous Christmas tree weeder
(ACW).
- The technical requirements to an ACW were defined to cover most plantation scenarios but
not extreme cases. Also the workings tasks for the first generation ACW were defined to
weeding only, as the development of an appropriate controller system would be the first
priority.
- The general developments within sensing technology, computing technology and robotics
has reached a stage where it is considered realistic to develop autonomous machines with
appropriate behaviour for regular agricultural and forestry field operations over longer
periods.
- An analysis of different ACW design outlines and operation patterns revealed that the
most suitable system was a small machine being able to move in between the trees in a
simple operation pattern relative to the location of trees, which are provided in a GIS
database.
- A suitable general type of system architecture was defined using sensor information on
position, orientation, surroundings structure, vehicle structure and self-awareness.
- The environmental benefits of using an autonomous machine in the described way were identified as:
effective weed control without use of herbicides, higher biodiversity because of left over weeds between the rows reduction of nutrient leaching and water erosion reduction of energy consumption.
- The overall annual costs of weeding with a future ACW is estimated to be similar to or
less than the present costs of mechanical weeding. The costs are more sensitive to changes
in the operation costs than to changes in the fixed costs, which means that the purchase
price of an autonomous weeder is less important.
- Added values can be achieved by means of data from the vehicle sensors or additional sensors on tree conditions for management decision purposes, by fitting auxiliary equipment, which can provide individual treatments of the trees.