ISCF WAVE 1 AGRI TECH: Robochick: an automonous platform for data-collection in poultry sheds

该项目的总体目标是证明商业家禽棚数据收集的移动的机器人系统的可行性。技术方法是首先建立系统的功能规范，用于动物福利友好型部署，然后调整Ross Robotics Ltd（RRL）模块化机器人平台，使其能够在人口稠密的棚内进行准确的自主导航和环境感知。将对选定的一系列传感器进行校准，并将其纳入一个多传感器模块，以将不同的环境参数融入小气候质量指数（空气和垃圾状况）。最后，我们将评估集成的移动的数据收集系统在商业环境中对关键的生产指标，并确定相关性，应存在与其他福利传感系统，如Fancom Eyenamic。这一技术方法与最近欧盟PLF项目采用的“活实验室”方法类似。应用家禽有限公司（APL）的商业设施代表了本可行性研究中的生活实验室。该项目将通过以下工作包运行：WP 1：鸟-机器人交互研究。将使用RVC的设施进行科学研究，以确定机器人操作参数和福利指标之间的联系，包括：发声水平;惊吓反应（移动距离，跳跃/翅膀拍打）;沮丧和警惕（不愿移动，摇头）;恢复正常行为的延迟。这一基本工作将从M1-M7运行。在M7中，该WP将提供一份报告，概述商业部署期间机器人的方法和功能规范。此次交付也是该项目的第一个里程碑; WP 2：物理平台和传感器集成的适应。在此期间，RRL的移动的机器人模块化平台将被优化，用于家禽棚内的操作。这将包括路线规划和动态避障方法的评估以及传感器模块设计和集成。将提供两个研究原型，一个在项目开始时用于鸟类-机器人交互研究，另一个在M10进行商业试验。后一个交付成果构成了该项目的第二个里程碑; WP 3：微气候多传感器开发。一个关键的目标是开发一个环境测量系统，能够准确地评估鸟类占领区的小气候质量。在对动物感知进行全面的水平扫描审查之后，将对环境传感器的准确性、响应、漂移、价格点和物理稳健性进行评估。传感器将包括温度、湿度和风速测量、用于粉尘测量的颗粒计数器以及用于CO2、NH3和VOC测量的传感器。初步调查将使用基于IRT的传感方法对鸟类的身体状况进行。该WP将提供传感器模块集成的最新审查和功能规范; WP 4：商业影响和可行性调查结果。从M10到M15，将对机器人对APL家禽棚商业生产的影响进行科学研究。机器人在棚内的操作将由熟练的畜牧工人使用基于天花板的视觉监控系统进行监控。将根据关键生产变量（例如，死亡率%、剔除率%、活重趋势、FCR、EPEF）评估机器人的影响。来自机器人的数据将与来自Eyenamic（Fancom bv）的数据和饲养员的专家观点相关联。将根据系统的能力和限制对RoboChick的数据价值进行评估，并提交商业评估报告（可交付成果）; WP 5：商业案例研究。可交付成果包括开发路线图文档、早期开发报告和最终设想系统的经济性报告。将在HAU举行一次宣传活动。

The overall objective of this project is to demonstrate the feasibility of a mobile robotic system for data collection in commercial poultry sheds. The technical approach is to first establish the functional specifications of the system for animal welfare-friendly deployment, then to adapt the Ross Robotics Ltd (RRL) modular robotic platform such that it is capable of accurate autonomous navigation and environmental sensing in a populated shed. A select range of sensors will be calibrated and integrated into a multi-sensor module to fuse different environmental parameters into indices of microclimate quality (both aerial and litter condition). Finally we will evaluate the integrated mobile data-collection system in a commercial setting against key production indicators and identify correlations that should exist with other welfare sensing systems, e.g. Fancom Eyenamic. This technical approach follows a similar modus operandi as recent EU PLF projects that use the "living lab" methodology. The Applied Poultry Ltd (APL) commercial facilities represents the living lab in this feasibility study. The project will run through the following work packages:WP1: Bird-robot interaction study. A scientific study will be carried out using facilities at the RVC to define the link between the robot operating parameters and welfare indicators including: vocalisations level; startle response (distance moved, jumping/wing flapping); frustration and vigilance (reluctance to move, head shaking); latency to return to normal behaviour. This essential work will run from M1-M7. In M7 this WP will deliver a report outlining the methodology and functional specifications for the robot during commercial deployment. This delivery also forms the first milestone of the project;WP2: Adaptation of physical platform and sensor integration. During this WP RRL's mobile robotic modular platform will be optimised for operation within a poultry shed. This will include the evaluation of approaches for route-planning and dynamic obstacle avoidance as well as sensor module design and integration. Two research prototypes will be available, one at the start of the project for the bird-robot interaction studies and one by M10 for commercial trials. This latter deliverable forms the second milestone of the project;WP3: Microclimate multi-sensor development. A key objective is to develop an environmental measurement system capable of accurately evaluating the microclimate quality in the bird occupied zone. After a comprehensive horizon scanning review of animal sensing, environmental sensors are to be evaluated for accuracy, response, drift, price-point and physical robustness. Sensors will include temperature, humidity and anemometry, particle counters for dust measurement and sensors for CO2, NH3 and VOC measurement. Initial investigations into birds' physical condition will be carried using an IRT-based sensing approach. This WP will deliver a state-of-the-art review and functional specifications for sensor module integration;WP4: Commercial impact and Feasibility Findings. From M10 to M15, a scientific study will be carried out on the impact of the robot on commercial production using APL's poultry sheds. The operation of the robot within the shed will be monitored by a skilled stockman using a ceiling-based vision monitoring system. The impact of the robot will be assessed against key production variables (e.g. % mortality, % culled, live-weight trend, FCR, EPEF). The data from the robot will be correlated with data from Eyenamic (Fancom bv) and the stockman's expert viewpoint. An assessment of RoboChick's data value measured against the system's capacity and constraints will be conducted and a commercial evaluation report (deliverable);WP5: Business Case study. The deliverables are an exploitation roadmap document, an early stage exploitation report, and a report on the economics of the final envisaged system. A dissemination event will be held at HAU.