In Sub- Saharan Africa, more than 60% of farms are managed by humans and less than 20% are motor-powered, a model that is not sustainable as food demand increases due to population growth of 11 million per year.
In addition, 35% to 50% of post-harvest losses for perishable agricultural products are lost annually in the region due to poor planting practices, according to the Food and Agriculture Organisation.
In 2014, Hello Tractor launched a mobile platform to allow farmers access to on-demand tractor services. The service aggregates tractor service requests (e.g. ploughing) using a mobile app and then pairs them with recommended tractors and operators while simultaneously tracking the number of hours each piece of equipment is serviced in the field and area.
Jehiel Oliver, CEO and Founder of Hello Tractor, said “Through valued relationships with companies like John Deere, we’ve been very successful in increasing mechanization access in smallholder communities. To reach the next level, we need to add additional services including predictive fleet utilization and maintenance; operator and tractor scoring; financing and the crop yield forecasting.”
Hello Tractor turned to the IBM research laboratory in Nairobi, Kenya in order to achieve this. Laboratory scientists work with Hello Tractor developers to apply a variety of technologies, including the Watson Agriculture, Blockchain, IoT and Cloud Decision Platform, to bring new services to the app for tractor owners and tractors, farmers and banks. The new services will be tested from the first half of 2019 in a pilot.
In particular, machine learning helps farmers to predict crop yields, which can be exploited in combination with advanced analytics and the blockchain to develop a loan credit score.
Weather data foreseen by The Weather Company, an IBM company; remote sensing data (e.g. satellite); and IoT data from tractors will also be included in the app to help smallholder farmers know when to grow, the quality of their farm cultivation, what to plant and the appropriate fertilizer using remote sensing and IoT information.
The IBM AgroPad technology developed at the IBM laboratory in Brazil could also be used to determine soil quality in the future.
Tractor Fleet Owners: Machine learning and IoT owners can view and manage the use of the fleet, predictive maintenance and forecast future use of tractors based on history, weather and remote satellite sensing data.
Tractor operators are classified and used on the basis of their training (e.g. ploughing, deep ripping, harrowing, fertilizing) using a five- star rating system. Owners will also have access to financing, maintenance and the purchase of new tractors and tools using historical data.
Tractor Dealers: After sales support, spare part inventory planning and credit administration can benefit from improved tractor repair and maintenance.
Banks and Financial Institutions: Can view and track utilization of tractors to determine a credit portfolio for the farmer and tractor owner, while also evaluating forecasted utilization to make credit decisions for tractor owners based on verified and trusted data on the blockchain.
Governments can use data and actionable insights to support different decision – making capabilities, such as to structure incentives, enforce investment priority regulations and policy decisions.
The backbone of the digital agriculture wallet is a blockchain and AI decision support platform that allows data to be captured, tracked and immediately shared while creating end-to-end trust and transparency for all stakeholders throughout the agribusiness value chain.
Dr. Solomon Assefa, Vice President, Emerging Market Solutions and Director, IBM Research – Africa, said “Our vision is to leverage AI, blockchain and the Internet of Things to digitize, optimize, and streamline agricultural business processes to create efficiencies and new services from farm-to-fork around the world.”
IBM researchers and Hello Tractor engineers intend to use machine learning with image recognition to predict the quality of cultivation for the next phase of the project. For example, remote sensing data can be used in combination with weather data to predict the next harvest.
Plans are also ongoing to extend the platform to Kenya, Mozambique, Senegal, Tanzania, Pakistan and Bangladesh beyond Nigeria.