Green Robot Machinery : Precision Cotton Picking Machine

Cotton is one of the largest commercial crop in India grown in 12 million hectares, which is 1/3rd of the world’s land under cotton. Cotton farming is the livelihood of about 6million farmers in India. 83% of the farmers own less than 2 hectares land and are classified as small farmers. Only 17% of the farmers hold 2 hectares and more.

Compared to cereals like wheat and paddy, cotton provides better income to the farmers, fetching 3 times the income compared to paddy farming. Cotton also requires lot less water compared to paddy and thrives well on good sunlight which, in India, is abundant throughout the year.

Cotton being a good drought resistant crop, it forms taproot, is grown almost all parts of India, however 70% of the crop is grown comes from the states of Gujarat, Maharashtra, Telangana and Andhra Pradesh. Recently states like Odisha , an eastern coastal state and Punjab the western states are promoting cotton as an alternative to paddy. Cotton is seen as a very good alternative to paddy, a water guzzler, both in terms of economics and environment friendly sustainable crop.

Interestingly the west Africa trio nations of Benin Bakino Faso, Cote De-Ivore are fast adopting cotton due to their favourable climate conditions and favourable economics. Cotton , an eco-friendly and sustainable crop, is the best fibre for human clothing. Surprisingly synthetic fibres, from non-renewal sources, is now used in 80% of the textile industry. Cotton if produced in enough quantities can replace the synthetic fibre to a large extend and support the world’s Sustainable Development Goals. Africa as a continent can play a major role in this growth.

Cotton production process

Cotton is a 160 days crop and the picking is done during the final 60 days of the crop. Cotton, being an indeterminate crop, is picked every 15 to 20 days and typically there are 3 pickings during a season. Cotton picking is done manually, mostly by women, in India. A human can on an average pick about 25 kgs/day.

The yield of cotton varies widely and the variation is large between irrigated and rainfed crops. Though primarily a rain fed crop, more and more of the crop is now being promoted under irrigated conditions. The official average yield of seed cotton in India is about 1500 kgs/Hectare and lint cotton yield, which is 35% of the seed cotton, is 532kg/Hectare. Efforts are on to improve the lint cotton yield to world’s average of 1000kg/Hectare which is 2850 kgs/Hectare of seed cotton.

Challenges in Cotton Farming

The challenge to a farmer in India is the peak time labour demands during the picking season. Socio economic pressures has led to exodus of farm labour to other less drudgery and more economical options. Better education and economic growth of the country will only accelerate the migration of human work force away from farm labour.

About 1/3 of the world’s cotton is harvested by harvesting machines which can harvest what is termed as synchronous bloomed cotton. Synchronous bloomed cotton is achieved by spraying plant growth regulators (PGR), a hormone, to induce blooming of cotton simultaneously and then spray the plant with a defoliant to shed the leaves making the plant ready for harvesting with no leaves on it. The mechanical harvesters requires the leaves to be shed from the plant otherwise the leaves themselves will be a thrash content which is not desirable. To obtain economic yields as the regular multi-bloom cotton, the farmers have to seed 4 times the seed rate of the regular multi-bloom cotton. The input cost for synchronous bloom cotton, which is increased because of higher seed rate, additional PGR and defoliant, will not be affordable to small farmers in India. There are also risks of crop loss due to pests, rain etc which are mitigated in multi-bloom cotton. Moreover the mechanical harvesters are not economical for small farms at the moment. Efforts are on by the agro machinery companies to build small compact harvesters suitable for the Indian farming conditions

Manohar Sambandam, Green Robot Machinery founder

I am a cotton farmer, owning 12.5 acres of land in southern state of TamilNadu in India. I am semiconductor engineer by profession with a Masters Degree in Electronics and Communication. I was leading the engineering development of semiconductors chips for WIFI, micro controllers and other segments. I also had the experience of being an engineering head of a WIFI startup which got acquired by one of the largest multinational which became the largest supplier of WIFI chips to the industry.

The problems in cotton crop also exists in horticulture crops especially in harvesting. Most of the horticulture crops require multiple harvesting. Multiple harvesting requires precision and cognitive intelligence to detect, locate and pick the harvestable produce similar a human. I decided to focus on cotton picking, the way humans picks cotton. Green Robot Machinery was formed with the primary focus of Farm Robotics for a sustainable agriculture. In the longer term the machine will be repurposed for other crops like okra, eggplant, tomatoes etc which require precision picking.

Many of the challenges relating to low yield in India are related to the inability of the farmers to intervene in a timely manner to address the issues of fertigation, weeding, pesticide spraying etc due to non-availability of timely labour. Robotics can address all these issue and also help improve the farm economics in a sustainable manner.

Green Robot Machinery autonomous cotton picker

Green Robot Machinery cotton picker consists of a multiple autonomous robotic arms mounted on a semi-autonomous electric farm vehicle. The robotic arm includes a stereo camera and a 3 degree of freedom actuators with an end effector for picking cotton. The camera can detect and locate cotton within 3mm precision in realtime. The actuator similar to a stripper dislodges the cotton boll from the cotton shell and a vacuum mechanism moves the cotton through the arm to a collection bin. The electric vehicle can move over a row of crop autonomously and the arm mounted on it does the tasks of picking.

The arm is targeted to pick about 50kgs/day and 4 of these arms mounted on the vehicle can pick about 200kgs a day, the average yield per picking, in a cotton farm in India. Higher yielding farms can be supported by adding additional arms.

The machine will be operated by a women which provide higher farm worker productivity, mitigating the critical labour challenge looming in the cotton farming in India. The machine is positioned as a productivity tool rather than displacing existing farm labour. In the longer term, if the situation demands, be more autonomy will be adopted for the vehicle. The machine also supports a sprayer as an add-on , which will help the farmers to do timely intervention of spraying of pesticide, herbicide and foliar spray, making the machine operational throughout the complete cotton crop season.

The challenges in precision farm mechanisation is first the productivity, robustness and reliability and finally the cost.

The machine if it can pick consistently 200kgs/day which is about 50,000 bolls, burst seed cotton, assuming a nominal weight of 4 grams it will meet the requirements of small and medium farmers. With 4 robotic arms doing this job has to perform 12500 repeated tasks of picking without any breakdown. Assuming a 10 hours run time of the machine per day the machine has to do about 20 pickings per minute which is the target for our robotic arm. The challenges is in the robustness and reliability of the parts in severe field conditions of humidity , temperature and dust conditions.

The cost of this machine will decide the economic viability of the solution on a commercial basis. From that perspective the machine is using all off the shelf standard parts from automobile, drone and EV ecosystem to keep the cost down and also ensure that components are available from multiple sources. The custom part includes metal fabrication of the housing and the 3D printed plastics of the robotic arm, which will be brought down once we go into the manufacturing phase.