| Front page | | Contents | | Previous | | Next |

Autonomous weeders for Christmas tree plantations - a feasibility study

9 Estimated performance of selected robot concepts

The Royal Veterinary and Agricultural University, Section for AgroTechnology

Having proposed possible outlines of autonomous Christmas tree weeders in the previous chapter the purpose of the following is to estimate the performance characteristics of these outlines.

9.1 Environmental benefits

Compared to the present practice of herbicide application the ACW weeding method has several advantages. Weeds can be controlled without use of herbicides and some natural plant cover may be left over between the rows to improve the biodiversity – an improvement that may help decreasing problems of pests and diseases, and thus the use of pesticides to control these.

Compared to mechanical inter row weeding the method also has the advantages that the soil surface is left nearly untouched and covered by plants, which will reduce problems of nutrient leaching and soil erosion. In addition the energy consumption and CO₂ emission will be much less, because cutting requires less energy than tillage, because only part of the area is treated, and a lightweight vehicle is used. The low weight will also reduce soil compaction.

9.2 Effectiveness of weed control

As described in chapter 4 weeding of 40 cm radius circular areas around each tree is enough to achieve good tree development. Both of the autonomous machine designs outlined in chapter 8 will weed more than this minimum area as they have to mow continuously between the trees. But still they will leave about 33 % (parallel operation pattern) and about 49% (loop operation pattern) of the area untreated. The weeds left in these areas may affect the tree growth positively as they provide shelter and increase biodiversity as already mentioned.

However, problems may occur from large weed plants or trees developing in the untreated areas. This may occasionally require supplementary treatments.

With the precision mechanisms used it is considered possible to weed within a few centimetres of the individual trees, leaving probably less than 1 dm² untouched. Weed development in these areas is not likely to cause essential problems because of strong competition from the trees.

9.3 Machine caused losses

Machine caused losses may arise from traffic of vehicles and from injuries caused by weeding tools.

Ignoring any possible negative effects of spray chemicals the only cause of damage associated to spraying may be mechanical injuries along spraying tracks. Also, the damage caused by the present type of mechanical weeding implements is little when they are properly used. It is estimated, that an ACW could reach the same low level.

However when it comes to headlands the present normal size machinery needs a rather broad area for turning, and because of the many turnings required, headlands are often not utilised or suffer from considerable damage. This problem would be solved or would become much less of a problem by use of a small autonomous machine, which can move between the trees at headlands.

9.4 Safety

The methods of achieving safe operation of the suggested ACW in relation to people, animals, trees and the machine itself (described in appendix A and B) are considered sufficient, as the machine is to work in an enclosed environment, is small and is to use screened working tools.

9.5 Estimated area capacity

The term area capacity is defined as the area treated within a unit of time. The area capacity of the beetle type ACW concept was calculated by use of typical or estimated data for the loop type and parallel operation patterns (figure 8.2 and 8.3). It should be noted, that the calculation refers to a design, where the rotor cutter is placed 20 cm off centre.

The parameter values and results are shown in table 9.1. It appears that the effective area capacity (capacity in the defined practical conditions) of the parallel operation pattern is 56% higher than for the loop type. If the off centre distances is reduced to zero the parallel operation pattern is superior by only 4 %, as the vehicle then gets a shorter travelling distance, but this is not considered feasible when the tree becomes some years old, because of the low branches on the trees.

Table 9.1.
Anticipated parameter values and corresponding area capacity of the selected types of ACW´s and operation patterns.

The sensitivity of the results to changes in the assumed parameters values are shown in figure 9.1.

Figure 9.1.
Sensitivity of the area capacity of weeding using an ACW in loop operation pattern.

From this it can be seen, that an increase of the forward speed and the row distance have high positive influence on the capacity. The same is the case for the field efficiency, which is not shown. Increases of the off centre distance and the cutting width have less and negative effects.

In case of the parallel operation pattern the field capacity will be proportional to the forward speed, and the field efficiency, while the other parameters have no essential effects.

9.6 Estimated costs

The level of costs associated with operation of an ACW was estimated using a calculation model and best judgement parameter estimates for different operation patterns, (table 9.2). It is of cause not possible to know how realistic these values would be some years a head, but to take an example a conventional self-propelled lawn mower of a suitable size would cost around 25000 to 35000 DKK. Transforming it into an ACW would require sensing systems and controls, which probably will cost considerably more than that, when buying general-purpose equipment. However, prices are expected to drop when the equipment is adapted to the specific purposes, and the electronic processes and software are compressed onto a few computer chips. As an example a general purpose RTK GPS, which now costs about 150000 DKK, is estimated to cost 10000-20000 when adapted to a specific purpose and a general base station. Also the general drop in prices of electronic hardware over time is likely to continue.

The cost estimate in table 9.2 shows an annual cost of around 3300 and 2400 DKK/ha respectively for the two operation modes when weeding 20 ha. This value should be compared to the current contractor rates, which are 1000 – 1500 DKK/ha for spraying of herbicides and 2500 – 3000 for mechanical weeding.

To get an impression of the relative influence on the specific costs of the various parameters a sensitivity analyses was made. For both operation patterns the analysis shows (figure 9.2), that the most important parameters are the area capacity, the parameters of the group of variable costs, and the number of annual weeding operations, while the purchase price and the area allocated to a single ACW are less important.

Table 9.2.
Duration and costs associated with the use of an autonomous weeder: A case calculation bases on anticipated parameter values.

Figure 9.2.
Sensitivity of the annual costs of weeding one ha using an ACW in the loop operation pattern.
   

Figure 9.3.
Sensitivity of the annual costs of weeding one ha using an ACW in one-sided "strait" operation pattern.

There may be different opinions about whether the chosen parameter values are realistic ore not, but at least the calculations provide an idea about levels and relative significance of the parameters. It appear to be important to find designs and operation patterns that provide high area capacity and low variable costs, while the ACW purchase price is less important. It also seems to show that competitiveness in relation to the present mechanical weeding machinery may be reached within a few years

9.7 Added values

The use of ACW´s in Christmas tree plantations opens possibilities of using the machine sensors and additional sensors to obtain information on individual trees in the plantation. This could for example be size, shape, colour, and occurrence of large weeds near to tree trunks and between the rows. This information could be used for improvement of management decisions like:

	The optimal time of growth regulation (partial debarking or cutting of lower branches), which may vary across the field,
	Supplementary weeding of large weeds,
	Spatial and single tree graduated application of fertilizer and pesticides, Placement of fertilizers and pesticides close to the trees, e.g. carried out by the ACW, to get maximum effect and minimum leaching.

In that way the consumption of auxiliary constituents can be optimised and most likely reduced on plantation basis. At the same time systematic data collection will enable a growth regulation in time and optimise the quality of the plantation.

The value of these opportunities could be quite considerable and would improve the overall economy of the machine.

9.10 Sub conclusions

It is estimated that the proposed ACW system compared to present mechanical weeding may deliver a better quality of work and considerable environmental advantages. On top of that there could be added values in terms of tree specific data for management decisions and costumer information. The costs of the system are estimated to be similar to present contractor rates of mechanical weeding.

| Front page | | Contents | | Previous | | Next | | Top |