AMSOS 360 Operational Group: Innovative 360-degree solutions for the sustainability of a new milk production system in farms with automated milking

• Access and download of robot data for each farm.
• Connect different sources of information in a relational database and establish an automatic cross-validation process.
• Develop a 360° smart management application that manages the information received to define alerts and send them to farm technicians to assess their level of significance and the urgency of action.
• Define new traits that express the adaptability of animals to the demands of the robot and animal genotyping.

Project Outcome 1 is called "Project Implementation and Monitoring," which includes activities related to coordination among the requesting members of the project's activities, project management, including modifications made throughout the project, and auditing payment requests. During coordination meetings, it is important that all participants in the operational group are informed about the project's progress, both in the activities in which they are directly involved and in other activities. This project has seen changes in the regulations regarding project justification over the different periods, so a significant portion of the coordination time has been spent reviewing the regulations and adapting documents, posters, and other justification materials to the newly issued regulations. In conclusion, we believe that in this type of project, it is important to properly quantify these items in terms of hours employed, considering all the personnel required for proper management, especially the justification section and the possibility of even facing additional audits with a very high cost in terms of labor hours.

Project Outcome 2 is called "Access and download of robot data from each farm," to enable access to and download of robot data from participating farms. To this end, the manufacturers of the most popular milking robots among CONAFE members were contacted. Contact was maintained with them to determine the conditions for access to the milking robots, the definition of the type of access, and the protocol to be followed for each farm to connect to the CONAFE server and download the information stored in the cloud or in the database managed by the farm software. The destination of this data was discussed, and an agreement was reached with two of the manufacturers, Lely and DeLaval. The agreement with the former requires payment and a signed contract with the operational member FEFRICALE, while the latter is free of charge. The Horizon (Lely) program connects via an API, while the DeLaval software connects to the CONAFE database via LogMeIn access. Farms' access permission is granted through the app in the case of Lely, while permission is given in writing in the case of DeLaval. At the end of the project, 40 farms were connected with DeLaval robots and 267 with Lely robots. This result resulted in difficulties with access to GEA robots. There are unclear boundaries regarding data ownership between the technology developers and the farms whose animals generate the information.

Project Outcome 3 , entitled "Adaptation of the hoof health and mastitis data collection program to enable animal identification using a portable RFID reader," consists of the acquisition of portable RFID readers and NFC phones to be made available to podiatry and milk quality technicians to facilitate the collection of lameness and mastitis data. The devices must allow for the automatic identification of cows for which lameness and mastitis information is collected. To this end, the data collection software is modified to incorporate RFID and NFC protocols. Many difficulties are encountered because there is no uniformity in the information transmission protocols in the farm animal identification devices developed by different manufacturers and in the first period of the project, in the midst of a crisis in the availability of computer chips, there were no RFID reading devices available on the market, so the project was modified to contract the connection service for the data collection applications to a company that had a reader available with which this task could be fulfilled, Geriontrading SL.

Project Outcome 4 is called "Connecting different information sources in a relational database and establishing an automatic cross-validation process." This process includes the design of the relational database and the flow of information from the different sources, the integration of all the information from the robotic farms with the historical information available at CONAFE and information from the collection of foot health and mastitis data, and the development of the database in SQL code. This is a five-phase process: access to information; matching and coding; validation; connection and standardization; and integration. The necessary equipment for managing all the information is also acquired, from RFID readers and NFC phones to tablets that install the software for collecting information on lameness and mastitis. A cow behavior monitoring system is also installed on a project farm to collect information on various reproduction, health, and nutrition parameters. This activity also includes recording lameness and mastitis data, which is complemented by information about the project at the farms visited. The total data collected is 57,000 foot health records from 216 farms and 5,600 udder health records from 94 farms. This activity has encountered obstacles due to the limited availability of readers on the market, which has necessitated modifying the project. It is worth mentioning that reading animal identification using the NFC protocol requires close contact with the animal identification device, which must be carried out with extreme care to avoid accidents. All equipment used is also exposed to dirt and potential impacts due to the characteristics of the environment in which it is used and the activities carried out.

Project Outcome 5 is called “Develop a 360° smart management application that manages the information received to define alerts and send them to farm technicians”, in which the initial criteria for establishing lameness and mastitis alerts are defined based on the information available on each farm and the history of each animal, a 360° smart management application is designed that analyzes the information that is incorporated daily into the database, connects it with the historical information of each animal, to detect foot health and udder health alerts for each animal, the application is tested and the results are checked with technicians and farmers. Finally, a software development is carried out to create an application for access to information by foot health technicians, udder health technicians and farmers to the information downloaded from the robot, with individual access keys according to the information to which access is granted. The ability to select the livestock, data period, individual animals for consultation and download of production data from individual milkings, milking flow, conductivity, milk quality parameters, cell count, daily quality parameters and others.

Project Outcome 6 is called “Establish a detailed report   with all the animal's historical information sent to the technician each time there is an alert, to assess its level of significance and the urgency of action and, if appropriate, reach a decision with the farmer. The report that the technician receives each time there is a lameness or mastitis alert is designed, as well as the protocol so that the alerts generated by the different cases defined using the current and historical data of each animal reach the corresponding foot health or udder health technician, who then takes the appropriate measures when returning to the farm and acting on the animal for which the alert has been received. The protocol follows an outline so that the daily processes of analyzing historical and daily animal data generate alerts that will be notified to the foot health technician through the DATPAT application and to the udder health technicians by email. When the podiatrist starts the application, all pending notifications to be reviewed are automatically downloaded from the server and are indicated on the notifications button (red or green button). By pressing The notifications icon opens the notifications window, where pending notifications appear by default, although all notifications sent can be selected and filtered by herd. When the podiatrist responds to a notification, they must select it and enter the date of assistance and the comment on the incident resolution. Milk quality technicians receive the alert by email at the same time as the herds, as currently there is no udder health information collection system similar to the DATPAT for foot health.

Project Outcome 7 is called "Define new traits that express the adaptability of animals to the demands of the robot and animal genotyping." A total of 6,600 cow samples were collected and genotyped in the laboratory. The following new traits were defined: Milk production/cow/day (24 hours): corresponds to the sum of the production of all milkings performed on the same day; Productive efficiency (kg of milk/min): defined by dividing milk production by the total time the cow spends in the milking stall; Fat percentage: average fat percentage of all milkings of the day; Protein percentage:   Average protein percentage of all milkings of the day; Average maximum milk flow during the milkings of the day; Total time in box: time elapsed from electronic identification of the animal until it leaves the box; Dead time: time taken by the milking robot to find the location of each teat and position the teat cups; Electrical conductivity of milk (in milliSiemens/centimeter); Front teat distance (FTD): calculated by subtracting the X-axis coordinate of the right front teat tip from the X-axis coordinate of the left front teat tip; Rear teat distance (RTD): calculated by subtracting the X-axis coordinate value of the right rear teat tip from the X-axis coordinate of the left rear teat tip; Udder depth (UD): obtained from the average of the Z-axis coordinate values of the 4 teat tips to express the average distance between the cow's teat tips and the ground; Udder balance (UB): defined as the difference between the mean Z-axis coordinates for the front teat tips and the rear teat tips; Front-to-rear udder distance (FRU): calculated as the difference between the mean Y-axis coordinates for the front and rear teats on each side. These traits must have sufficient genetic variability to be susceptible to genetic selection.

Project Outcome 8 is called "Implementation of a genetic and genomic evaluation for new traits and establishment of a total economic merit index for robot farms." This involves implementing a genomic evaluation for thirteen robot traits using the 'Single Step' procedure. This method allows for the joint evaluation of genotyped and non-genotyped animals using the kinship matrices obtained from the pedigree and the genomic kinship matrix using the genotypes performed on the project farms. The calculated genetic indices correspond to genomic values that are combined into a total economic merit index. The economic merit index is used to select animals based on their total genetic merit while maximizing benefits due to their higher production performance, adaptability to the robot, and lower costs. The estimated economic values for this index are: kg milk, €0.22/kg/day; kg fat, €1.35/kg/day; kg protein, €2.89/kg/day; % fat, €12.57/tenth; % protein, €29.83/tenth; Productive life, €0.14/day; Days open, €6.96/day; SL (RCS), €34.08/unit; Clinical mastitis, €63.03/cow/lactation; Milking speed, €29.48/cow; Calving ease, €36.81/cow; Milking flow rate, €0.0142/kg/min; Box time, €0.095/min; Infectious lameness, €57.2/cow/lactation.