In exercise and rehabilitation, to effectively train the human body, human motion trajectory is essential because it induces muscle activity patterns. In this letter, we develop a novel framework for the trajectory optimization of human postures, including the head, the limbs, and the body to induce patterns of target muscle activities. Our framework has the following features: 1) a data-driven muscle-skeleton model for managing user-specific features; 2) quaternion-based state representation amenable for IMU sensors in human posture measurement; 3) joint optimization of human postures to replicate therapists who adjust not only paralyzed limbs but also patient's other limbs and body postures. We experimentally investigated the effectiveness of our framework with a shoulder joint assistive exoskeleton robot for rehabilitation.

中文翻译：

---

基于四元数的人体姿势轨迹优化诱导目标肌肉激活模式

在运动和康复中，为了有效地训练人体，人体运动轨迹是必不可少的，因为它会诱发肌肉活动模式。在这封信中，我们开发了一种新的框架，用于人体姿势的轨迹优化，包括头部、四肢和身体，以诱导目标肌肉活动的模式。我们的框架具有以下特点：1）数据驱动的肌肉骨骼模型，用于管理用户特定的特征；2) 基于四元数的状态表示适用于人体姿态测量中的 IMU 传感器；3) 人体姿势的联合优化，以复制治疗师，不仅可以调整瘫痪的肢体，还可以调整患者的其他肢体和身体姿势。我们通过实验研究了我们的框架与肩关节辅助外骨骼机器人进行康复的有效性。