Otto Motors是用于在制造和仓库运营中自动化材料处理的自动移动机器人（AMR）的制造商。这些机器人旨在以自由移动的方式将材料带到人，设备，架子和搁置。借助内置的协作功能，AMR是优化工作流程的高产方式。
According to the company, Otto robots have the ability to double productivity, but they have to be installed correctly in the first place. Simply using a spreadsheet to figure out how many AMRs are needed in an operation comes with a cost if there are too few—or too many—robots. Even a 20% variable on a fleet of 100 AMR vehicles means that either a customer is not getting what they need or are spending more than they should to achieve what is necessary, which makes for a weaker ROI.
但是有一个解决这个问题的方法，它以仿真软件的形式出现。
Matthew Rendall, CEO and co-founder of Otto Motors, was at the机器人峰会和博览会5月，在波士顿发表了有关使用模拟成功部署的演讲。
“我们相信模拟是一个重要的先决条件in your toolkit for a large fleet implementation,” Rendall said. “And the reason is that you can be highly granular, highly variable, and very catered to the specifics of not only the layout of the facility, but how your operation changes day by day, shift by shift, and minute by minute. And then you can load all of those parameters into a simulation and model it for a full operation.”
It's basically a way to start with the end in mind and mitigate the risk of failure via scenario analysis. The larger the fleet, the bigger the risk. “If you have 100 AMRs on a line, you are probably integrated into mission critical business processes, “Rendall explained. “And if that system doesn’t work the plant doesn’t work, and if the plant doesn’t work that is catastrophic. So simulation is important for risk mitigation.”
Rendall went on to describe two real customer scenarios, one greenfield deployment and one brownfield. They have some similarities, but in a greenfield project—where you are not dealing with existing infrastructure—the benefits of operational efficiency are very high.
AMR仿真
The greenfield site was a 1 million sq. ft. facility covering production and warehousing. It had 300 pickup/dropoff locations and AMRs would do 5,000 deliveries and cover over 1,000 miles per day. Because they were starting from scratch, Otto was able to partner with the customer to help them rethink how their manufacturing facility of the future would be laid out. Then they could simulate scenarios testing for routing, traffic, location, etc.
模拟分为两类，物理模拟和过程仿真，将物理模拟的输出用作输入过程仿真。这是有趣的事情发生的地方。
“We put the robots into the matrix,” Rendall said. “They don’t know they’re in a simulation as the software is fully emulating the environment. So we teleport the vehicle into the matrix and it’s operating with no knowledge it’s in a simulated environment versus a real environment.”
From there, they can conduct micro-tests to look at interactions and understand how much time an AMR will spend waiting at intersections and pickup locations, for example. These things are validated in the real world and feedback is created in a closed loop micro-simulation to ultimately look at the entire end-to-end process to see how the operation unfolds.
 |
有关移动机器人的更多信息：AI和机器人车队管理。 |
在这种情况下，制造商实际上减少了所需的AMR数量，同时仅通过能够通过每种情况分析来提高效率。Rendall解释说：“它最初是在102 AMR上范围的，但是通过模拟，我们能够将设计优化为83 AMR。”因此，尽管有些客户可能会想到要在模拟软件上花费更多的前期，但另一方面确实有真正的价值。“在模拟上进行一些投资，并通过更智能的设计和更有信心节省30％的车队规模。”
布朗菲尔德模拟也采用类似的方法，基本差异是基于已经发生的决策的自由度较低，但是该设施已经存在。它也经常处理包括叉车，AMR和人在内的交通混合。因此，这里的方法更多是关于如何在不中断操作的情况下将AMR置于AMR中，以最大程度地减少停机时间。
In either scenario, however, the use of simulation results in the same outcome. “Our position is that a small amount of investment upfront is so important in order to set the project up for success and to make sure you are spending the right amount of money to solve the right problems and mitigate the execution risk associated with going live with an AMR implementation,” Rendall said.
