Robotics in Agriculture

Seminar Report
Robotics in Agriculture

GUIDED BY:
Dr. D.R.Kathiriya
Sign: __________

SUBMITTED BY:
Deepak Rajpurohit
(06-0173-2015)

Page 1 of 10

Robotics in Agriculture

CONTENTS
Introduction.................................................................................................................... 3
Seed bed preparation..................................................................................................... 3
Seed mapping................................................................................................................ 4
Seed placement.............................................................................................................. 4
Reseeding....................................................................................................................... 4
Crop care........................................................................................................................... 5
Weed mapping............................................................................................................... 6
Robotic weeding............................................................................................................. 6
Micro spraying................................................................................................................ 7
Fruit picking robot:.......................................................................................................... 7
Selective harvesting....................................................................................................... 8
Advantages Of Robotics in Agriculture............................................................................9
disAdvantages Of Robotics in Agriculture.......................................................................9
Conclusions.................................................................................................................... 9
References.................................................................................................................... 10

INTRODUCTION
Page 2 of 10

Robotics in Agriculture

The idea of robotic agriculture (agricultural environments serviced by smart machines) is not
a new one. Many engineers have developed driverless tractors in the past but they have not
been successful as they did not have the ability to embrace the complexity of the real world.
Most of them assumed an industrial style of farming where everything was known before

hand and the machines could work entirely in predefined ways much like a production line.
The approach is now to develop smarter machines that are intelligent enough to work in an
unmodified or semi natural environment. These machines do not have to be intelligent in the
way we see people as intelligent but must exhibit sensible behaviour in recognised contexts.
The approach of treating crop and soil selectively according to their needs by small
autonomous machines is the natural next step in the development of Precision Farming (PF)
as it reduces the field scale right down to the individual plant.
Automatic sensing and control (on-the-go) for each task is also important and many research
papers have shown that these systems are feasible but most are too slow, and hence not
economically viable, to be operated on a manned tractor.

SEED BED PREPARATION

Ploughing is one of the most important primary cultivation processes and has been carried out
since the start of civilization. It is effectively the inversion or mixing of topsoil to prepare a
suitable seed bed. It also has the ability to bury surface crop residues and control weeds. A
small robot utilising current technology does not have the energy density to sustain ploughing
over a large area due to the high levels of energy needed to cut and invert the dense soil.
Secondly, the draft force required to plough also needs relatively high weight to give traction.
Perhaps we would leave it at that, but by considering what the plant, or in this case the seed
actually needs, we can approach the problem in a different way. The seed requires contact
with the soil moisture to allow uptake of water and nutrients, it requires stability to hold the
growing plant and a structure that allows the roots to develop and the shoots to grow. A
solution is two fold. Firstly if we do not compact the soil in the first place there is less need
for energy inputs for remedial loosening. Natural soil flora and fauna can be encouraged to
manipulate the soil to give a good structure. This is one of the reasons to opt for smaller
Page 3 of 10

Robotics in Agriculture

machines. Secondly, if the majority of the soil rooting depth is acceptable, then only the local
environment of the seed needs to be conditioned before seed placement, which will take a lot
less power. Add to this the ability to place nutrients in the correct proximity to the seed we
can improve the early phase of establishment. This system has many of the advantages of
direct drilling but incorporation of previous crop residues may still cause a problem although
removal of crop residues is an option.
SEED MAPPING

Seed mapping is the concept of passively recording the geospatial position of each seed as it
goes into the ground. It is relatively simple in practice as an RTK GPS is fitted to the seeder
and infrared sensors mounted below the seed chute. As the seed drops, it cuts the infrared
beam and triggers a data logger that records the position and orientation of the seeder. A
simple kinematic model can then calculate the actual seed position. The seed coordinates can
then be used to target subsequent plant based operations.
SEED PL ACEMENT

Rather than just record the position of each seed it would be better to be able to control the
seed position. This would allow not only allow the spatial variance of seed density to be
changed but also have the ability to alter the seeding pattern. Most seeds are dropped at high
densities within each row, whilst having relatively more space between the rows. From first
agronomic principles, each plant should have equal access to spatial resources of air, light,
ground moisture, etc. Perhaps a hexagonal or triangular seeding pattern might be more
efficient in this context.
RESEEDING

Reseeding is the concept of being able to identify where a seed was not planted, or that a crop
plant has not emerged and a machine can automatically place another seed in the same

position. This concept could be extended to transplanting a seedling instead of a seed if the
surrounding plants are too far advanced. A reseeder would have the ability to insert individual
seeds/plants without disturbing the surrounding crop. Conventional seeders could not then be
used as they create continuous slots in the soil. A punch planter could be developed to fulfil
Page 4 of 10

Robotics in Agriculture

this role, or better still adapt a Japanese transplanter to deal with one seedling at a time. Prior
local micro-cultivation could be achieved by using a targeted water jet (or gel) to pierce the
soil and soften it ready for
the seedling roots. Figure 1, show a transplanter adapted to take a seeding mat. The seeding
mat can also include crop nutrients. If this concept became efficient enough, it could also
become the main seeder as well.

Figure 1. (Left) Japanese transplanter adapted to take a seed mat, (Centre)

CROP CARE

Crop scouting

One of the main operations within good management is the ability to collect timely and
accurate information. Quantified data has tended to be expensive and sampling costs can
quickly out weights the benefits of spatially variable management. Data collection would be
less expensive and timelier if an automated system could remain in the crop carrying a range
of sensors to assess crop health and status. A high clearance platform is needed to carry
instruments above the crop canopy and utilise GPS. Smaller sub canopy machines have been
developed in student competitions. Examples of both types of machines are shown in Figure
2.

Page 5 of 10

Robotics in Agriculture

Figure 2. (Left) Portal crop scouting platform (Right) Sub canopy robot
WEED MAPPING

Weed mapping is process of recording the position and preferably the density (biomass) of
different weed species using aspects of machine vision. One method is to just record the
increased leaf area found in weedy areas as weeds are patchy and the crops are planted in
rows (Pedersen 2001). Another more accurate method is to use active shape recognition,
originally developed to recognise human faces, to classify weed species by the shape of their
outline. Current research has shown that up to 19 species can be recognised in this way.
Colour segmentation has also shown to be useful in weed recognition. The final result is a
weed map that can be further interpreted into a treatment map.
ROBOTIC WEEDING

Knowing the position and severity of the weeds there are many methods that can kill, remove
or retard these unwanted plants. Different physical methods can be used that rely on physical
interaction with the weeds. A classic example is to break the soil and root interface by tillage
and promote wilting of the weed plants. This can be achieved in the inter row area easily by
using classical spring or duck foot tines. Intra row weeding is more difficult as it requires the
position of the crop plant to be known so that the end effector can be steered away. Within the
close -to-crop area, tillage cannot be used as any disturbance to the soil is likely to damage

Page 6 of 10

Robotics in Agriculture

the interface between the crop and the soil. Non contact methods are being developed such as
laser treatments and micro-spraying.

Figure 3. (Left) The autonomous Christmas tree weeder, (Right)

MICRO SPRAYING

Within the close-to- crop area, great care must be taken not to damage the crop nor disturb
the soil. One method of killing weeds close to the crop plants is to use a micro spray that
delivers very small amounts directly on to the weed leaf. Machine vision can be used to
identify the position of an individual weed plant and a set of nozzles mounted close together
can squirt a herbicide on to the weed.

FRUIT PICKING ROBOT:

The principles of fruit picking robots have been developed since the early 1980s. These
principles have opened up new approaches to the harvesting of crops. However, to fully
develop the fruit picking robotics technology, contributions from high-tech industry,
agricultural commodity groups and farm equipment manufacturers must be sought.

Page 7 of 10

Robotics in Agriculture

To start with, the fruit picking robots need to pick ripe fruit without damaging the branches or
leaves of the tree. Mobility is a priority, and the robots must be able to access all areas of the
tree being harvested. It goes then without saying that the robots must be intelligent, and have
a human-like interaction with their surroundings through senses of touch, sight, and image
processing. The fundamental blocks of these robots are shown in the figure

SELECTIVE HARVESTING

Selective harvesting involves the concept of only harvesting those parts of the crop that meet
certain quality thresholds. It can be considered to be a type of pre-sorting based on sensory
perception. Examples are to only harvest barley below a fixed protein content or combine
grain that is dry enough (and leave the rest to dry out) or to select and harvest fruits and
vegetables that meet a size criteria. As these criteria often attract quality premiums, increased
economic returns could justify the additional sensing.
To be able to carry out selective harvesting effectively, two criteria are needed; the ability to
sense the quality factor before harvest and the ability to harvest the product of interest
without damaging the remaining crop. Most agricultural equipment is getting bigger and
hence not suited for this approach. Smaller more versatile selective harvesting equipment is
needed. Either the crop can be surveyed before harvest so that the information needed about
where the crop of interest is located, or that the harvester may have sensors mounted that can
ascertain the crop condition. The selective harvester can then harvest that crop that is ready,
while leaving the rest to mature, dry, or ripen etc.

Page 8 of 10

Robotics in Agriculture
ADVANTAGES OF ROBOTICS IN AGRICULTURE

The Robots does not get sick or tired and does not need time off.

It can operate with closer tolerances (so every round is at full field capacity), Fewer
errors and at higher speeds

Because machines can be made lighter and cheaper if the drivers seat, controls and
cab can be eliminated.

It can be used in various fields like agriculture, medicine, mining, and space research .

It can be sent to another planet to study their environmental conditions.

The machines could easily work around trees, rocks, ponds and other obstacles.

Small suburban fields could be worked almost as efficiently as large tracts of land.

DISADVANTAGES OF ROBOTICS IN AGRICULTURE

One of the key disadvantages of driverless machines for agriculture is liability.

Access to the technology.

Not currently scale neutral.

Better sensors would help. Improved scouting programs would be essential.

Robots could change the culture /emotional appeal of agriculture.

Energy issues ,costly.

CONCLUSIONS

One of the advantages of the smaller machines is that they may be more acceptable to the
non-farm community. The jobs in agriculture are a drag, dangerous, require intelligence and
quick, though highly repetitive decisions hence robots can be rightly substituted with human
operator. The higher quality products can be sensed by machines (colour, firmness, weight,

Page 9 of 10

Robotics in Agriculture

density, ripeness, size, shape) accurately. Robots can improve the quality of our lives but
there are disadvantages too.

REFERENCES

Page 10 of 10