We will work with the project end-users/stakeholders [farmers, environment protection and sustainable development experts, public agriculture authorities (local, national and also EU wide), veterinarians, cooperatives, livestock scientists (researchers) specialized distributors] to identify:

  • a bill of user needs, environmental and integration constraints, to help us prioritizing most impacting and beneficial developments
  • hindrances identified in the existing solutions
  • available technologies, components
  • applicable legislation, directives

    Beyond a clear and updated state of the art in the subject, carried out by the academic partner (INRAE), firstly, we will identify relevant stakeholders, coordinate a group of key end-users (already been identified during the preparation of the proposal); organize workshops and interviews with them, for then planning further field experimentation and measurements. This will allow us to define user needs and requirements and identify gaps and areas of potential improvement in the market offer (this has started before the project, and we have high-level knowledge of the needs, gaps and improvement areas). We will understand all about end users current methods and daily operations; technical solutions they currently use; integration and environmental constraints.

    The outcome will be a list of user requirements, classified by priority. Field experimentation and measurement campaigns will be necessary to select and fine tune the set of sensors that are able to provide the information collected and feed with those the analytical and predictive capabilities of the IoT and AI platform.

    The most relevant goal of the project is the proper and efficient application of data to the target stakeholders’ decision making process. Apart from improving current devices, we look forward to contributing to the existing processes applying differential machine learning techniques to interpret key aspects for extensive livestock husbandry such as geolocalization of the herd, social and individual behaviours, herd movement throughout the year and spatial resources allocation (i.e. allowing assessment of efficiency in the use of natural resources and landscapes), monitoring of essential physiological stages such as peaks of parturition (i.e. for matching e.g. nutritional requirmeents with grasslands surfaces rotation strategies), predation, possible links with health issues and environmental conditions, etc., from available sensors’ data. Data will be gathered from different measurement campaigns and pilots, and will feed the AI algorithms.

    Secondly we will map user needs, including environmental and integration constraints, into a system and service proposition that complies with requirements and constraints; a Functional Specification will be produced. The outcome will be a clear System and service architecture document and a Technical Specification for the three main system components (devices, user’s Smartphone application, web-based IoT Platform). This Technical Specification will lead development and integration work. Interface Control Document (ICD) will be part of the System and service architecture.

    A Functional Test plan will be delivered too. Thirdly, we will select, among all available technologies, those that can fill in the gaps and comply with the Technical Specification and the System and service architecture. We will generate a rationale of the technologies that will be applied to the project in order to respond to requirements, considering what is available from the market. Three key factors will lead component selection:
  • the addition of costs of every component keeps the bill of material (BOM) under the cost constraints imposed by the business plan and the pricing model for our target market (extensive livestock farmers with limited margins): the final solution is affordable.
  • the components and devices fulfil environmental constraints (IP protection level, temperature operation/storage range…) and integration constraints (size, shape, weight, power consumption…).
  • the components are scalable to mass production: there is no risk of obsolescence or provision shortage, or this can be properly managed; the long life of the components and low failure rates ensure an economically viable life and failure rate for the devices.

    Concerning algorithms and Artificial Intelligence, research will be carried out to list state of the art technologies and select the most adapted paradigms. Many technical meetings will be organized to coordinate WP1 work towards a coherent system architecture and specification.

Leave a Comment