摘要
以下引文由新闻编辑从发明人提供的背景信息中获得：“已经提出使用机器人模拟器来生成在机器学习G模型(一个或多个)的训练中使用的模拟训练数据。机器学习模型(一个或多个)一旦基于模拟训练数据训练，然后可以在实际机器人(多个)的机器人控制的感知、规划和/或动作阶段(多个)中使用(即，然而，在机器人控制中由真实的Rob OT(s)使用机器学习模型(s)并不主动地涉及机器人模拟器(s)。相反，机器学习模型(s)仅仅基于使用机器人模拟器(s)生成的模拟训练DATA来训练。在将控制堆栈实施到实际机器人上之前，已建议使用机器人模拟器（Robotic Sim ulator）来测试/验证机器人控制堆栈。然而，一旦在实际机器人上实施机器人控制堆栈，机器人模拟器（Robotic Simulator）就不再被积极利用。
Abstract
The following quote was obtained by the news editors from the background informa tion supplied by the inventors: “Use of robotic simulators has been proposed to generate simulated training data that is utilized in training of machine learnin g model(s). The machine learning model(s), once trained based on simulated train ing data, can then be utilized during perception, planning, and/or acting stage( s) of robotic control of real robot(s) (i.e., non-simulated physical real-world robot(s)). However, utilization of the machine learning model(s) by the real rob ot(s) in robotic control does not actively involve robotic simulator(s). Rather, the machine learning model(s) are merely trained based on simulated training da ta that is generated using the robotic simulator(s). Further, use of robotic sim ulator(s) has been proposed to test/validate robotic control stacks prior to imp lementation of the control stacks on real robots. However, once the robotic cont rol stacks are implemented on the real robots, the robotic simulator(s) are no l onger actively utilized.”
NETL
NSTL