Role of Nonuniformly Quantized Actuation in Biological Motion Generation

该奖项的目标是建立一个生理启发的框架，招募符合生物运动生成的模块化执行器。该研究将深入了解神经肌肉系统的模块化和可变性如何在协调多块肌肉中发挥关键作用。 研究了柔性驱动器阵列的非均匀补充方法。 浮点量化（FPQ）是数字通信中流行的编号方案，它模仿了Henneman运动单位募集的大小原则。该研究假设：（1）肌肉力量和信号依赖性的变化可以通过非均匀模块化致动器系统中的量化致动曲线来表征，（2）量化致动器招募方法再现了这种信号依赖性的变化，而无需将人工噪声源引入机器人架构，以及（3）肌肉协调中的优化原理可以通过从随机控制的角度将个体肌肉水平上的可变性映射到终点处的可变性来数学建模。此外，还将开发用于生成生物轨迹的肌肉型机器人驱动器的量化控制方法，并将由此产生的人造机器人肢体的运动性能与人类的运动性能进行比较。如果成功，该研究将推动仿生机器人、康复机器人、计算神经科学和角色动画领域的发展。对神经肌肉生理学的更深入理解将提供更复杂的计算模型，从而产生一种新颖的机器人架构，该架构捕捉了生物运动系统固有的优势。研究成果将被整合到当前的课程，包括为本科设计研究提供一个特殊的主题。研究生和本科生将从跨学科和多元文化群体，包括代表性不足的群体招募。K-12学生的外展活动将通过机器人夏令营计划，格鲁吉亚技术的学生和教师增强伙伴关系计划，和第一机器人进行。

The goal of this award is to establish a physiologically inspired framework for recruiting compliant modular actuators for biological movement generation. The research will result in an in-depth understanding of how modularity and variability in the neuromuscular system play key roles in coordinating multiple muscles. The research studies a non-uniform recruitment method for compliant actuator arrays. Floating point quantization (FPQ), a popular numbering scheme in digital communications, appeared as a viable option which mimics aspects of Henneman's size principle of motor unit recruitment. The research hypothesizes that (1) variability in muscle forces and signal-dependency can be characterized by quantized actuation profiles in non-uniform modular actuator system, (2) a quantized actuator recruitment approach reproduces such signal-dependent variability without introducing an artificial source of noise into robotic architecture, and (3) optimization principles in muscle coordination can be modeled mathematically by mapping the variability at the level of individual muscles to the variability at an end-point from a stochastic control perspective. Quantized control methods for recruiting muscle-type robotic actuators for biological trajectory generation will be developed, and resultant motor performance in artificial robotic limbs will be compared to performance in humans.If successful, the research will advance the field of biologically inspired robotics, rehabilitation robotics, computational neuroscience, and character animation. A deeper understanding of neuromuscular physiology will provide more sophisticated computational models, resulting in a novel architecture for robotics that captures the advantages inherent in biological motor systems. Research outcomes will be integrated into current courses including offering a special topic for undergraduate design research. Graduate and undergraduate students will be recruited from interdisciplinary and multicultural groups including under-represented groups. Outreach activities for K-12 students will be conducted through a Robotics Summer Camp program, Georgia Tech the Student and Teacher Enhancement Partnership Program, and FIRST Robotics.