Robotics in agriculture : a (r)evolution ?

Agriculture is a demanding sector, where the work is often difficult and requires the use of a variety of tools and machines. The nature of the land to be treated is also varied, with conditions that can change greatly depending on the weather or humidity. The wide variety of activities to be dealt with, and the difficulty of the work involved, make agricultural tasks arduous and even dangerous. Indeed, in addition to the immediate risks associated with the use of increasingly imposing machinery, the arduous nature of the work involved and the toxicity of the products used make this one of the activities most exposed to socio-professional risks. Nevertheless, agriculture is an important pillar of society, since it must be able to feed a growing world population. Subjected to ever-increasing economic constraints, recent awareness of our planet's vulnerability has added a new constraint on the environmental impact of human activities. Against this backdrop, agriculture needs to rethink its production system in order to combine economic efficiency, productivity and respect for the environment. Robotics is a promising way of resolving such paradoxes. The growing autonomy of robots means that they can now be used to operate in difficult areas, or to carry out tasks requiring the use of hazardous materials. In addition, the need to reduce the use of plant protection products is leading to an increase in the precision of treatments, and/or the mechanical performance of certain tasks requiring much longer working times. The introduction of robotized operations makes such a change conceivable, and leads to the vision of a new model of agricultural production. This major impact is anticipated in economic analyses concerning the development of the robotics market. A 2014 analysis ( https://www.tractica.com ) forecasts considerable revenue growth in this field, with a figure of $74 billion in 2024, compared with less than $3 billion in 2015, making it the second-largest robotics market. However, to reach these levels, robot performance must be compatible with the diversity of tasks to be carried out, as well as with the wide variation in work contexts. Climatic conditions and the different types of soil encountered throughout the year in the agricultural context lead to major disparities in the behavior of robots, which today perform well when interaction conditions are well mastered.