Integration of energetic cost into optimal control models of reaching

人们更喜欢以最小化能量的方式移动，但缺乏一种统一的方法来理解跨领域的首选运动。在运动方面，我们知道人类运动系统更喜欢能量消耗最小的行走方式，当新的力量施加在身体上时，它可以改变人们喜欢的行走方式，与能量消耗最小化相一致。在人类接触方面，该领域最近刚刚测量了每次接触所消耗的能量。首选的到达方式——比如手移动的路径、手的速度和肌肉的力量——现在被认为反映了能量消耗和任务奖励之间的权衡，当新的力量施加到手上时，首选的到达方式的变化也与最小化能量消耗相一致。然而，解释能量消耗在到达过程中的作用的生理和预测机制仍有待开发。在接下来5年的研究计划中，我的目标是确定能量消耗如何影响人类运动系统控制伸手行为的方式。在这项工作中，我将使用最优人类到达的计算模型来预测三个不同领域的首选行为和能量消耗数量。首先，我将确定能量成本模型如何预测正常到达的基本行为和成本（目标1）；接下来，我确定能量成本如何影响河段对外力的适应（目标2）；最后，我们将确定能源成本如何影响控制政策的发展（目标3）。在电机控制中，术语“控制策略”具有特定的含义，在这种情况下，指的是被控系统的状态与该状态的动作或施加的扭矩和扭矩之间的函数关系。我们用标记和摄像系统，或者机器人外骨骼来测量手臂的运动，我们用外骨骼施加力来改变运动的动态。为了测试我们的预测模型，我们测量了人们喜欢的运动的变化，以及他们消耗的能量，并将其与我们的模型进行比较，该模型预测了给定任务的位置、速度、扭矩、运动持续时间和相关的代谢成本。然后我们可以确定能量消耗和任务目标如何影响运动系统对到达的控制。提出的研究将提供一个基本的和基础的理解如何人类运动系统协调平稳，优美的运动。这在加拿大工业中具有广泛的应用意义，例如：了解对衰老和损伤的自然适应；基于能量成本最小化的合理康复方案；用于学习和控制的辅助技术和设备。我们的研究项目将为跨学科学生提供丰富的培训环境，为加拿大工业界和学术界的研究和技术做出贡献做好准备。

People prefer to move in ways that minimize energy, yet a unifying approach to understanding preferred movements across domains is lacking. In locomotion we know that the human motor system prefers ways of walking that have minimize energetic cost, and that it can change the way that people prefer to walk when novel forces are applied to the body, consistent with minimizing energetic cost. In human reaching, the field has just recently made measurements of the energy consumed per reach. Preferred ways of reaching-such as the path the hand travels, hand speeds, and muscle forces-are now thought to reflect a trade-off between energetic cost and task reward, and changes to preferred reaching when novel forces are applied to the hand are also consistent with minimizing energetic cost. However, a physiological and predictive mechanism that explains the role of energetic cost in reaching remains to be developed. Over the next 5 years of my research program I aim to determine how energetic cost affects how the human motor system controls reaching behaviour. In this work I will be using a computational model of optimal human reaching to predict both the preferred behavior, and quantity of energetic consumption, in three distinct areas. First I wil determine how the energetic cost model predicts fundamental behaviors and cost in normal reaching (Aim 1); next I determine how energetic cost affects the adaptation of reaches to external forces (Aim 2); and finally we will determine how energetic cost affects the development of control policies (Aim 3). The term `control policy' in motor control has a specific meaning, and in this case refers to a functional relationship between state of the controlled system, and the action or applied torques and torques for that satte. We measure the motion of the arms using either a marker and camera system, or a robotic exoskeleton, and we apply forces using that exoskeleton to change the dynamics of movement. To test our predictive model we measure changes in people's preferred motions, and the energy they consume, and compare that to our model, which predicts positions, velocities, torques, movement durations, and the associated metabolic cost, for a given task. We can then determine how both energetic cost and task objectives affect the motor system's control of reaching. The proposed research will provide a basic and foundational understanding of how the human motor system coordinates smooth, graceful motion. This has broad applied implications in Canadian industry, such as: understanding natural adaptations to ageing and injury; rationally based rehabilitation programs accounting for energetic cost minimization; assistive technologies and devices for learning and control. Our research program will provide a rich training environment for interdisciplinary students, well prepared to contribute to Canadian research and technology in industry and academia.