https://thedairylandinitiative.vetmed.wisc.edu/home/housing-module/adult-cow-housing/automated-milking-systems/

Check out this excellent resource for AMS facility design by The Dairyland Initiative! 

Always good to remember that we need to understand and consider cows' behavior when setting goals for efficiency on farms!

The main take-home messages for me are:

• Design for 55 per robot and a minimum of 2 AMS units per pen: Cows are not robots, and since the desire to be milked is not constant throughout the day, it is highly unlikely that cows will ever maintain the constant throughput through the robot required to achieve the theoretical robot capacity.
• Provide a minimum of 24"(60 cm) of feed bunk space per cow with frequent push-ups: As shared in this article, I have also seen multiple AMS farms with 3-4 row pens with limited bunk space. I often hear that since the cows have extra time on their budget and do not have a "fix schedule", the limited bunk space would not be an issue (since not all cows eat simultaneously). I am not familiar with any data that confirm these assumptions in the literature, and therefore, I agree that we should stick with what we know about cows eating patterns.
• Functional Gating: "One of the main reasons for switching to AMS is labor efficiency. It is absolutely essential that AMS facilities be designed for one person to move cattle easily and efficiently around the robot with the least amount of stress".
• Lameness rates are higher in AMS herds than in conventional parlor herds: That was surprising to me since I usually hear farmers that transitioned to AMS saying that after the transitioning period (in which cows from the conventional barn are introduced to AMS) and lame cows are culled or removed from the system, they usually perceive fewer cows getting lame, likely because they need to walk less and have more time to rest. It appears that, although there is potential to reduce lameness in AMS (because of the factors I just mentioned), decisions on design and management (use of mattress stalls, slatted floors, inadequate hoof care, and use of footbaths) can trigger the opposite effect.
• Depending on the AMS facility design, footbaths are used less frequently because it is difficult and time-consuming. Often, footbaths are located in crossovers, and cows are only foot bathed on bedding day - a practice the Dairyland Initiative does not recommend. Footbaths should be designed and located so that cows are automatically sorted through the footbath when exiting the robot and/or returning to the resting area. Easy access to a trim chute for inspection of a lame cow's foot should also be a priority.

When adopting new technology and designing a new facility (usually not very cheap), it is easy to overlook the cow's needs for logistics/convenience and cost sake. Although this can be an attractive choice, it usually costs much money in the long run. Do not be afraid to take the time to visit as many operations as you can and talk to as many people (with as many different perspectives as you can get) before starting the project.

If you are thinking about transitioning to AMS and have questions, need help with facilities design (you can also contact Tim Terry - txt2@cornell.edu), or are looking for any resources, do not hesitate to contact Camila Lage (cd546@cornell.edu or 607-422-6788).