arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2207.13409cs.ROcs.SYeess.SY
收藏

经典控制中的无偏主动推断

Unbiased Active Inference for Classical Control

TU Delft(代尔夫特理工大学);Oxford Robotics Institute(牛津机器人研究所);Oxford University(牛津大学)

Mohamed Baioumy, Corrado Pezzato, Riccardo Ferrari, Nick Hawes

摘要:

主动推理是一个数学框架,起源于计算神经科学。最近,它已被证明是在机器人技术中构建目标驱动行为的一种有前途的方法。具体来说,主动推理控制器(AIC)已经成功地在几个连续控制和状态估计任务。尽管它相对成功,一些既定的设计选择导致了一些实际的限制机器人控制。这些包括对状态的有偏估计,以及只有一个隐式的控制行为模型。在本文中,我们强调这些限制,并提出了一个扩展版本的无偏主动推理控制器(u-AIC)。u-AIC保留了AIC的所有令人信服的优点,并消除了其局限性。在2自由度机械臂上的仿真结果和在7自由度机械臂上的实验结果表明,u-AIC相对于标准AIC具有更好的性能。代码可以在https://github.com/cpezzato/unbiased_aic上找到。

英文摘要:

Active inference is a mathematical framework that originated in computational neuroscience. Recently, it has been demonstrated as a promising approach for constructing goal-driven behavior in robotics. Specifically, the active inference controller (AIC) has been successful on several continuous control and state-estimation tasks. Despite its relative success, some established design choices lead to a number of practical limitations for robot control. These include having a biased estimate of the state, and only an implicit model of control actions. In this paper, we highlight these limitations and propose an extended version of the unbiased active inference controller (u-AIC). The u-AIC maintains all the compelling benefits of the AIC and removes its limitations. Simulation results on a 2-DOF arm and experiments on a real 7-DOF manipulator show the improved performance of the u-AIC with respect to the standard AIC. The code can be found at https://github.com/cpezzato/unbiased_aic.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑