engineering Archives - Agri-EPI Centre

Tag Archive for: engineering

Collaboration essential for successful agri-robotics

By: Eliot Dixon, Head of Engineering at Agri-EPI Centre

Robotics has several strong applications in agriculture, especially in scenarios where systems can enhance the productivity of a shrinking workforce or can offer production efficiencies to the farm. However, to be successful in these applications the systems created need be reliable, in terms of long-term physical robustness but also in the ability of their control software to handle the very wide variety of scenarios they will encounter in a farming environment. This means the robots must be both well designed and well tested to meet the needs of farmers. This includes a design which emphasises safety and reliability.

“Understanding user requirements and testing in-field is key”

Good design requires a deep understanding of the needs and requirements of farmers and their farming systems. This extends from the core values held by a farmer, such as safety, which dictate their decisions; through to very specific requirements created by the unique combination of their way of working and the land they work. If this understanding is not achieved for a farming system, then there is a very high chance that the eventual product will be unsuitable, either creating a failed product or a long development timeline to solve the deficiencies. Gaining this understanding should come through working with a wide variety of farms within the target market for the technology, not just a small handful. In many agricultural sectors this design stage is especially important due to the limited testing season and ability to iterate on the design.

Testing is also well understood to be important to creating a reliable product, and in agriculture this does require a close collaboration with farmers to ensure that the robot meets their needs. As these are complex machines, which are also often dangerous if not created with a strong safety process, the testing regime should also be rigorous enough to ensure that the system will function to the desired reliability for all the design requirements. A rigorous testing regime would usually require multiple tests for each requirement across multiple operational scenarios such as different weather conditions, soil types, dangers, failure modes, crops etc. Failure to complete this testing will certainly result in the robotic system encountering situations which it is unable to function within, which may create unfortunate repercussions for the user or manufacturer. Unfortunately, completing this massive number of tests requires a range of test facilities, some of which might be beyond the capability of a company focussing on a small range of agricultural applications.

In our 2021 hackathon we explore safety and security. Outcomes are discussed in our white paper here:

Hackathon white paper

As mentioned, good design and testing is essential to creating successful products, but this unfortunately comes with a high cost. Doing this for the wide range of complex operating scenarios in UK agriculture, as well as the short testing cycles, is driving up the cost of developing agricultural robots. There are a multitude of Agri-robotics companies in the UK creating their systems from almost the ground up, each of which are individually bearing the cost in time and money of this development. This creates barriers to adoption in terms of high costs, a limited set of operations which can be conducted by robots, or low reliability due to poor engineering, and is increasing the amount of time it takes for products to get to market. As in all development the saying “Good, Cheap, Fast. Pick two”, is very much in action here but some very pressing needs mean we must find ways to break that deadlock.

Collaboration enables future opportunities for robotic systems

The obvious solution for this deadlock is to massively increase collaboration between ag-robotics developers. This has been proposed for many years, but we are yet to see a viable solution to this. Direct collaboration is currently difficult for commercial reasons with developers competing for the same money, but also for technical reasons where it is challenging to share components between robots. Perhaps a solution for this is to build an ecosystem of adaptable, compatible, components and platforms which can be used to create a multitude of agricultural robotic systems. This ecosystem of components would also be able to be robustly tested to ensure reliability when integrated as part of a larger system. Thus, the costs of development would be increasingly shared, without any single robotics manufacturer losing income as they are all developing for specific agricultural niches. Using a set of well proven components would allow developers to focus on ensuring good understanding and design for specific problems in agriculture, while also allowing for easier integration and testing of the robots.

Robotics in agriculture is a promising field, and with the right design and testing, as well as collaboration between developers, it could be a great success. By understanding the needs and requirements of farmers and using that to create an ecosystem of components and platforms, robots can be developed which are high value, robust, reliable and safe. With the right approach, agricultural robotics could benefit farmers across the UK and worldwide. Read our robotics and automation article to understand more about how we can support you to develop a robust well tested solution through collaborative R&D today.

Agri-EPI expands robotics and data offering

Agri-EPI has developed its robotics and data offering, including the addition of 4 new members to their engineering team over the last couple of months.

Eliot Dixon, Head of Agri-Tech (Engineering) explains:

“Over the last few months Agri-EPI has been investing heavily in its engineering team, bringing on several new members, enabling us to offer a set of services to assist in the creation of agri-tech products. The team of platform and spectral imaging experts uses our fleet of sensors and specialist software to deliver a range of sensing products such as ground truthing for AI model generation, or the creation of digital twins. We are also now able to offer UAV and UGV platforms as a means to test novel sensors and end-effectors without the need for a bespoke vehicle. And through working closely with our innovation farm network, we are creating a heavily layered source of evidence for developers using our farm network to design and test their innovations.”

Agri-EPI’s new GIS Data Analyst, Yingwang Gao, majored in Agricultural Engineering, and has a PhD degree specialising in Hyperspectral Imaging Applications, as well as postdoc experience working as Research Associate. In addition to a strong academic background, he has accumulated several years of industrial work experience, mainly on spectral imaging systems, R&D, and spectral imaging data analysis in various application domains. He has a strong passion for remote sensing and photogrammetry. At Agri-EPI, he takes care of data acquisition and data processing from different types of sensors, including RGB, multispectral, hyperspectral, LiDAR, and GPR, to identify and map out features of interest in the agricultural sector, to help farmers with better decision-making in agricultural management.

Agri-EPI’s new R&D Equipment Technician, Aditya Jadhav, pursued his bachelors in aeronautical engineering, where he learned various aspects of flying machines. He set up an aeromodelling club with a few of his classmates where they designed, built and tested various configurations of small UAVs. The MSc program for Autonomous vehicle dynamics and control was structured for students to gain a deeper understanding of unmanned systems. Aditya was part of a group project that built a surveillance system with a swarm of autonomous drones, and an individual project sponsored by the Railway Safety and Standards Board which aimed to design and develop an autonomous vehicle which can operate in a station environment. The advancements in robotics and the urgent need of integrating robotics with sustainable agriculture were the driving forces for him deciding to work in the agri-tech sector. As the R&D Equipment Technician, Aditya looks after all the deployable assets that are in service to the company, which includes maintenance, asset tracking and deployment, and organising the logistics.

Panagis Tzivras, Agri-EPI’S new GIS Software Engineer, is a GIS expert with strong technical skills who is highly invested in programming. In his previous roles working with startups and the commercial sector, he was involved in data collection and extraction, maintaining data pipelines and building geospatial processes and automation updates. At Agri-EPI Centre he is helping to leverage the measurement resources of the centre to create high quality dataset and support systems. He is working on creating tools and code to enable the automation of data collection from a wide variety of sources available to Agri-EPI Centre.

Lastly, Aidan Robertson has joined the Agri-EPI Engineering team as their new Graduate Data Analyst. Aidan’s background is in mathematics, which he studied at University of Warwick for four years before looking for jobs related to data science. He is very pleased to be part of the Agri-EPI team in quite a varied role; so far, he has been working on projects related to the health and wellbeing of cows, specifically by reformatting farm datasets to be sent out for analysis. Soon, there are plans for him to begin a more ambitious project to develop a costings estimator for RAS in agriculture. This is a long-term task, but the ultimate goal would be to offer it as a service for farmers looking to introduce robotic systems into their farms. The most interesting part of agri-tech for Aidan is the data, and what it actually says about the performance of a system, as well as what can be done to help the problems being faced by the agri-tech sector at present.

R&D in Automation and Robotics for agriculture

By: Eliot Dixon

Agri-EPI, the centre for precision innovation in farming, is a first choice for agri-tech developers, from start-ups right through to established companies, to help with creating robust and commercially viable agricultural solutions.

I am Eliot Dixon, the Head of Engineering at Agri-EPI. I have a technical background in automotive engineering which has taught me the importance of good systems engineering,  but also am lucky enough to be part of a family with an active farming business. These dual backgrounds have shown me that it is vital that agri-tech solutions are built on well described initial design goals created from a strong understanding of the needs of farmers and their operations. If that is not done, then there will be delays in the development of the product and eventually quality, which will have ongoing negative effects on the trust of farmers in the product.

This is especially necessary in agricultural robotics, where highly complex technical systems and operating environments coupled with a very short testing season gives very little room for mistakes or iteration.

Precision innovation aim

Our aim is to guide developers through the innovation process of understanding their design requirements and creating and testing their new technology. This ensures that farmers gain access to profitable and productive solutions to empower more sustainable farming.

Our offer

As a company we do this through a wide-ranging set of facilities, equipment, and services that cover data, spectral imaging, data analysis and modelling, real world testing facilities and robotic platforms for validation and trials.

Key resources include:

  • Academic links with leading agri-robotics universities
  • A commercial farm network to develop system requirements and conduct in-field testing
  • Project management
  • Consortia building
  • Development services and equipment services for developers
  • Delivery Team

My technical background is in intelligent robotics, enabling robotics to understand and react to their environment, which I see as a key component in a robust agri-robotics system. The offering of the team and wider organisation is shaped by this to enable us to deliver many of the needs of developers working with intelligent robotics.

Our team is a resource that can be accessed as a service for any UK organisation who would like to join us in a commercial or research collaboration. We help in the development process through a combination of a strong team and a world class set of equipment and facilities.

My team is made of specialists from multiple technical domains. Between us we have academic backgrounds in ground robotics, aerial robotics, computer science, physics, mathematics and spectral imaging, and have employment experience in academia, defence, automotive, aerospace, agri-tech and manufacturing. The engineering team works as part of the wider technical team, delivering on our promise of development support from ideation right through to commercialisation.

Our farm network is a key part of this, enabling the testing spaces and long-term interaction with farmers which we rely upon. The team also works outside of the farm network with our deployable equipment, which is the major topic of this article. We will take a closer look at our farm network data offering in a future article.

Whilst I’m very proud of the skills of the team, we do also have an extremely exciting set of resources at our disposal which we are very keen to share. When looking at this from a robotics point of view, our services can broadly be split into two categories: platforms and sensors. Both sets of services are operated from our hub at Cranfield University.

At Agri-EPI we see the need to develop a UGV or UAV platform for a specific agri-tech product as something which slows down development of new applications of those technologies. Therefore, we have invested in manufacturer-independent development platforms which allow sensors and end-effectors to be created without needing to create a bespoke system or work directly with a platform developer. This allows collaboration with platform providers to happen only when the requirements of the sensors/end-effectors are fully understood. Our most interesting offers here are our UGVs, Sam and Frodo, and our multi-purpose UAV platform. These can be quickly adapted to almost any agricultural scenario and have the onboard processing power to unlock their full capabilities as a platform. Members of the team have extensive experience working with platforms such as these.

We are also aware that some sensing technologies which might be extremely useful for robotics development, especially in the domain of spectral imaging, are a very large investment in terms of equipment cost and personnel, and can be  difficult for developers to justify even if the returns can be large. For this reason, we continue to invest in our sensing capabilities and our ability to analyse that data, and we share that resource as a common capability for UK Agri-Tech. We provide high quality sensing across a broad range of technologies, including hyperspectral, SIF imaging, multi-spectral, ground penetrating radar and LiDAR. Almost all these sensors are airborne and are useful for creating data sets used in machine learning training, agronomy, simulations, and system validation. They are particularly useful for the arable domain, but we can modify the way we deploy them for most other agricultural domains.

For both services (platforms, and sensors) we offer a service provision from creation of the initial testing plans right through to a delivery of analysed data. Planning of operations is conducted in-house, especially in the case of our UAV mounted systems, and we also undertake post-processing of sensor data using the spectral imaging expertise of the team and a suite of industry leading software.

If you are an agri-tech developer who has a particular interest in robotics, or you require assistance in using some difficult sensors, then we would love to hear from you. Get in touch here or fill out this form.