Robotics in agriculture: the future is already here
Lorenzo Marconi of the Department of Electrical and Information Engineering of the University of Bologna, an expert in automatic controls and founder of the university's spin-off company FieldRobotics - which develops robotic systems capable of operating in hostile environments - tells Mondo Macchina about the current landscape and prospects of robotics applied to agriculture.
Over the past few years, you have developed a UGV, a rover with interchangeable equipment for specialized tasks which, after having been managed as an internal DEI project, is now followed by a full-fledged company. Would you tell us about the evolution of the project and what are the specific features of your robot?
Work on the UGV began more than five years ago. We decided to invest in the development of a ground vehicle with precision agriculture in mind, transferring a whole series of skills that we had honed in some European robotics projects, especially on the subject of autonomous navigation in hostile environments, and making use of the specific mechanical engineering skills gained in the group. The area of agriculture immediately seemed to us to be the one where the impact we could bring was greatest. The project started more as an academic research activity, an attempt to develop wireline navigation algorithms without the use of GPS, only based on real-time reconstruction of the surrounding environment with cameras and laser scanners. Many undergraduates, postgraduates and PhD students took turns on the prototype and contributed to improving it, attracted by the field of automation in agriculture that increasingly attracts the interest of young engineers. As the years went by, we saw that the vehicle became more and more reliable, even operating in very hostile environments such as terrain, and with some mechanical devices it could perform many operations that made it a "universal" machine. So we decided to invest and gamble on bringing it closer to the market and to an industrialization phase. FieldRobotics, a spin-off of the University of Bologna, was born with the aim of making this type of rover the tractor of the future, the tool through which the field work will be combined with instrumental mapping to carry out applications in precision agriculture.
There is no doubt that in recent times there has been an increasing amount of talk about robotics in the agricultural field, both in relation to the advantages deriving from the precision of the solutions proposed, and for the possibility of reducing the risk for human beings to zero, especially when working in extreme conditions. What is your vision?
FieldRobotics' vision is that the roboticisation and electrification of the field is an inevitable process that the entire world of agriculture will necessarily have to converge on. The trend we are seeing in the automotive sector, in terms of vehicle electrification and the development of autonomous driving, will also characterize the agricultural sector and it is strategic to invest in this direction. There is definitely an issue of job security, but also economic sustainability. The agricultural sector suffers more than others from the problem of finding both skilled and unskilled labour. The variability of production from one season to the next makes it increasingly complicated to organize work teams. In addition to the logistical aspect, there is clearly the economic one which, in the field of fruit growing for example, makes manual harvesting no longer sustainable. Our goal is to bring the revolution we witnessed a few decades ago in other areas of production, such as manufacturing, to agriculture as well. The difficulties are enormous, since agriculture is not as structured as packaging or manufacturing. It's a revolution, it's going to take time, but we're convinced it's going to happen. And the revolution will involve not only machines but also everything behind them, such as service and the mindset and skills of the users.
Do you think the future of these applications will be "electric" in terms of power generation and machine movement?
I really think so and not only for a question of economy of the electric system compared to endothermic systems and for a question of environmental sustainability, but also because the electric actuation and movement allows to obtain much more compact mechanical designs, which are more reliable than those in which the motion is generated by an endothermic engine. I think we will see more and more decentralization of motorization, that is, many peripheral electric motors positioned where the motion is needed. This follows by a few decades what has happened in the field of industrial automation where mechanics has slowly been replaced by mechatronics. Not to mention the whole world of greenhouse production where electrical implementation is essential for obvious reasons. This electric revolution, similar to what's happening in automotive, will also bring with it the need to upgrade the infrastructure with, possibly, on-site generation systems using renewable sources.
For some years now, France has been leading the way in Europe in the field of orchard and vineyard robotics, with several companies already on the market with machines for almost every type of processing. Why is it that in Italy, in terms of industrialized products at least, we are so far behind in this sector? Are we in danger of having to chase the competition? What is missing in your opinion?
Italy, like other European countries, has a significant gap with France. For several years now, the French government has put an unimaginable amount of resources into stimulating innovation in agriculture, and a whole series of visionary industrial realities have taken advantage of this opportunity to gain a competitive advantage that is there for all to see. As noted at recent trade shows, Eima to name one, there are already commercial machines developed by French startups that allow them to perform tasks fully automatically that were unthinkable just a few years ago. We are certainly behind but, considering the resources that will arrive in the agritech area, for example through the Recovery Plan, I am sure that the gap, even if not completely closed, can at least be reduced. The important thing is not to waste time and to put into play all the imagination, expertise and technology that sees Italy excel at the international level. Indeed, I think it's just a matter of resources, certainly not skills and the ability to innovate.
Robotics and artificial intelligence are currently under the EU Commission's lens. The new Machinery Regulation will address these issues, a harmonized standard, the future EN ISO 18497, is being written that will define safety requirements for autonomous machines, and the Commission is writing a regulation regarding Artificial Intelligence. How important do you think it is for Italian companies to be active in these areas? What are the risks or issues associated with future regulations?
I think it is very important that Italian companies engaged in innovation in agriculture are strongly involved in this regulatory process. The problem of regulating aspects of artificial intelligence in agriculture, as well as in other areas, must necessarily be addressed step by step and through close collaboration between regulatory bodies, industrial partners and even universities. It's a domain where technical expertise is high and it's hard to think about identifying the right rules without strong technical involvement. When it comes to artificial intelligence, there can be many contexts and they cannot be homologated in a single instance. It is therefore necessary to first regulate simpler aspects, such as artificial intelligence aspects related to autonomous navigation in environments not frequented by humans, and then go on to regulate more complicated and uncertain aspects, such as those in which there is "co-habitation" between autonomous or semi-autonomous robotic systems and human beings.
These aspects are now standard practice in other areas of production and will also characterize the agricultural world in the future. The risks I see are related to overly stringent regulations, perhaps motivated by a lack of understanding of the specificities and characteristics of artificial intelligence algorithms, which could hinder development by making it more difficult to catch up with countries ahead of us like France.
In the race for robotics, we are currently seeing those targeting large autonomous tractors, those targeting smaller, more specialized machines, and even fleets of small collaborative rovers. Looking to the future, what do you think will be the evolution of robotics in the field of agricultural processing?
I think it really depends on the application context. There is no doubt that in certain areas of agriculture, large machines are unavoidable to ensure productive efficiency. FieldRobotics, which is more interested in fruit growing and greenhouse production, thinks the vision of having lots of small robots and managing large expanses and large productions by multiplying these small machines is a winning choice and definitely a futuristic scenario. Multiplying machines, while keeping them small in size, has a number of advantages.
For example, the problem of soil compaction is drastically reduced. In addition, the machine's smaller size ensures lower costs and easier maintenance. Redundancy also allows for greater flexibility in planning predictive maintenance without stopping production.
In the context of fruit farming, moreover, the small size enables a different exploitation of the land, with plants that can have smaller inter-row dimensions. FieldRobotics is gambling in this direction.
What is, and what will be, the role of UAVs in Agriculture?
So-called drones today are constrained by limited flight range and limited cargo capacity. Therefore, their use is mainly confined to the area of monitoring with limited weight sensors.
Their most obvious benefits are clearly associated with the potential speed at which data are acquired (a few minutes are needed to map a hectare) and the ability to have privileged vantage points. I think their prospective use is always in synergy with ground vehicles complementing their characteristics.
Personally, i see UAVs as useful to determine health maps quickly, with mapping frequencies much higher than satellite solutions, always in synergy with ground vehicles that could represent docking stations for airborne agents in order to compensate for the limited autonomy. Also in this case the technologies available today are at a very high level both in terms of reliability and ability to operate in hostile environments.
However, their use is not always justified, especially in Italy where the average size of farms is few hectares. I still think you need to be ready by mastering these technologies. This is the case at FieldRobotics where many of the contributors have a lot of experience in this domain.
Autonomous robots compete in the field
Testing the behaviour and performance of autonomous robots through a competition aimed at verifying which machine is able to offer the best performance. This is the objective of the European project Metrics, which involves 17 institutional players from all over the continent - Italy is represented by the Politecnico di Milano, the Universities of Milan and Florence - and which involves four different fields of activity: agile production, health, infrastructure maintenance and agrifood. Each sector corresponds to a specific sub-project: the one concerning the agri-food sector is called ACRE (Agri-food Competition for Robot Evaluation). In this sub-project, field trials are designed to evaluate the ability of robots to perform tasks that require a high degree of decision-making autonomy from artificial intelligence, such as removing weeds or providing pesticide treatments.
Daedalus, the rover that does everything on its own
The use of autonomous robots for agricultural work is already a reality, and in the near future they are likely to become even more widespread. In the field of automation and artificial intelligence, progress is the order of the day, and it is happening at a rapid pace. The synergies between companies, universities and research centres, as well as targeted and substantial investments, are making it possible to develop increasingly advanced and precise AI systems, more accurate sensors, and more advanced and user-friendly software applications. It is already possible to see some of these machines at work in our companies, while others are at the prototype stage but are not far from making their market debut. This is what is happening with the development of Dedalo, a project born from the collaboration between the University of Bologna and two Bolognese companies, Fieldrobotics and Aslatech Robotics Company. Dedalo is an autonomous tracked rover, capable of carrying out various cultivation operations typical of fruit trees, such as chopping and providing pesticide treatments. Compact and light (550 kg without equipment), the robot is powered by a 12 kW battery with an autonomy of 9 hours (on level ground). Dedalo is able to move through spaces thanks to a technology that recognises obstacles and detects the size and structure of rows.
Robotic technologies, a challenge for the Italian system
Today there are around one hundred types of agricultural robot in the world, including machines that are actually operational - some of which were presented to the general public at EIMA 2021 - and prototypes still in the testing phase. A significant share of these machines, around 30%, is intended for operations such as weeding and harvesting. From a purely quantitative point of view, the record for latest-generation robotics is held by the USA, which, with 24 models developed by US companies, occupies first place in the world ranking. They are closely followed by the Netherlands and France, which can boast 16 and 12 autonomous robots respectively, ahead of Germany (7). Italy, which can boast four highly innovative robots, the same number as those developed by Great Britain, Spain and Denmark, is also in the running. The figures, quoted in a recent report published by Informatore Agrario, show how the Italian research system is committed to recovering competitiveness compared to other countries, and indicate how fundamental it is that this commitment be supported with programmes that involve public bodies and private companies in equal measure.
