The year in review: Take a bite out of 2021+

Everything We’ve Been Up To This Year

As 2021 draws to a close, it’s a time for reflection!

From ethical, legal and social implications of human-robot collaboration in industrial automation, use cases and user requirements, to real-life demonstration and validation of the CoRoSect system, here’s a glance at our year in review.

Ethical, Legal and Social Implications (ELSI) of human-robot collaboration in industrial automation

KU Leuven Centre of IT & IP Law (CiTiP) combines insights from experts in robotics, AI, computer science, and engineering with insights in philosophy, law, psychology, and sociology, to prevent ethical challenges. The Ethical and Legal Framework report offers a better understanding of the developed technologies. Released at the end of 2021, the report offered a list of challenges involved in the development of AI and robotics in insect farming and how to mitigate them. The work completed so far provides a basis for the evaluation of the platform and for the recommendation of policies to enhance a safe and healthy human-robot work environment.

Bridging the gap from user requirements to system architecture

Through collaborative work of the entire consortium, we defined 5 types of Use Cases. Over 3 months of analysis was performed to assess the business value and technical feasibility of every task to be performed by the CoRoSect system so that all stakeholders can get proper understanding of the tasks and challenges. Furthermore, we dedicated 3 months facilitating the Pilots definition process. During this process, we performed rapid interactions between the partners, such as one-on-one and collective topical sessions. The 5 Use Cases include: Quality Management and Intervention, Oviposition Management, Larvae/Cricket Management, Pupation Management, Harvesting and Waste Management and they are all tailored to the needs of the insect farms end users.

Biological, technical and economic aspects and parameters to insect rearing

What we want to achieve is to get insights into biological, technical and economic requirements of insect rearing processes and improve them through management, sensing and automation. The purpose of this work during the first year was to describe environmental constraints and biologically and economically optimal conditions for rearing of black soldier fly, mealworm and house cricket. The outcome is a set of parameters to be monitored and controlled when developing a decision support model and digitising insect farming which can vary from species to species. During 2021, the applicability of some of the sensor measurements have already been piloted at Luke’s InsectLab.

Farm-level modelling and orchestration

CoRoSect provided the description of the systems architecture that will be implemented and allow the automation of the insect farms. Concepts were analysed and different functionalities and commands have been identified to improve the management of all the smart components in the shop floor, thus defining the systems involved in the performance and achievement of the use cases. Ongoing work is focused on managing the operations carried out by the components in the plant simultaneously with developing an initial version of the IMS from the information collected by the technical partners. 

Boosting autonomy using AI-enabled perception methods

We focused on equipping each farm cell with advanced AI perception methods and made a lot of exquisite progress in the three following areas: environment analysis & registration, human motion analysis & prediction and integrating physics-based force feedback interactions in VR applications.

From human instructions to robotic actions

Here, the CoRoSect project’s focus was on planning, experimentation and initial development of tasks, specifically for the Versatile force-adaptive control for handling crates and insects, the SLAM/Navigation of AGVs and the implementation of safety control in robotic planning. We are now in the phase where we are making trials and experimenting with different alternatives.

Essential technology: cognitive robots and smart mechatronics

CoRoSect partners worked together on designing, manufacturing and testing novel sensors and electronics solutions, based on the use cases and user requirements collected, that can be integrated into the crates which support and improve the automated insect rearing process. At the end of the first project year, the initial prototype of an integrated intelligent sensor was created to test the functionality of the planned structure and way of implementing it. In addition to smart sensors, CoRoSect partners worked on developing the design of an autonomous guided vehicle (AGV) and two robot cells that will cooperate closely together. 

Speeding up innovation through human-robot collaboration schemes

The focus here has been on algorithms that let the robots learn to solve complex real-world tasks in insect farms by mimicking the natural way of learning in humans through visual observation. We researched relevant techniques, which were later tested using reinforcement learning with image-based information. Communication between ROS and Microsoft’s Hololens 2, which was achieved using ROS-TCP-Connector, was made to implement mechanisms that are necessary to facilitate human-machine interactions for situation awareness, as well as human worker training for human-robot collaboration-based insect care activities. 

Meeting the requirements of insect farming through secure platform integration 

The first year’s progress was towards identifying the different assets that will be integrated within this system, covering mainly the Shop Floor and the Management level, and towards working with system providers to define the integration protocols and the schemas for the digitalization of the devices. In addition, we held several workshops to introduce technical partners to these protocols who will assist in their adoption to get everything fully integrated. 

Working towards a real-life demonstration and validation of the CoRoSect system

Over the past couple of months we’ve been heavily working on the design of pilot scenarios to test and evaluate the CoRoSect’s system under real-life conditions. Under the guidance of our partners from TECNOVA, NASEKOMO and HSEL, the CoRoSect team carefully gathered information on the actual operations performed by insect farms, their infrastructure conditions, software components, hardware devices and data layer. Eventually, eight pilots were selected, two pilots in each insect farm: Italian Cricket Farm, Invertapro, Entocycle and Nasekomo. 

AgriFood Forum 2021: from ambition to action

Working side by side with AgriFood robots – opportunities and challenges

Under the umbrella of Agrifood Forum—one of the most important agrifood events in Europe, CoRoSect recently joined forces with its sister projects Robs4Crops, FlexiGroBots and the Robotics4EU CSA. Some of the industry-leading names touched upon the main drivers and barriers for automated agrifood futures. 

According to Dr. Rico Möckel, there is a massive push for robotization in various areas, and agriculture is one of them. Cognitive robots must be capable of carrying out complex tasks in dynamic environments. Ethics must be a priority when robots and humans work in collaboration—harnessing the potentials of AI and reducing its risk.