基于表面肌电信号的人体运动意图识别方法研究

提出了基于表面肌电信号进行人体运动意图识别的新方法。该方法将智能发育思想结合其中,分为离线知识库建立和在线知识库检索两个过程。首先,进行离线训练,建立IHDR(Incremental Hierarchical Discriminant Regression)知识库。采集人体表面肌电信号,对肌电信号进行小波分解,将各层小波系数的方差作为每个通道信号的特征;然后,将每个通道的特征作为输入,人体运动行为作为输出,离线建立基于智能发育思想的IHDR知识库。人体运动意图识别的过程即知识库的检索运用过程,通过每个通道的特征来在线检索IHDR知识库,搜索到合适的节点即为运动意图。实验表明,基于智能发育思想的人体运动意图识别方法可以达到满意的识别率。

智能发育方法研究及其在机器人行为规划问题中的应用

优化问题是机器人在动态非结构环境下进行决策的关键问题之一。但是，优化算法通常具有较高的计算时间复杂度，尤其是在包括多种动力学约束的情况下，这限制了其在实时系统中的应用。受到人类智能发育过程及最新神经科学研究发现的启示，本文提出一种基于智能发育机制的场景驱动控制结构，对人的知识积累和应用知识的过程进行模拟，研究了与该控制结构相关的关键技术，包括智能发育算法、在线训练、知识的积累和检索方法等，并将这种控制结构应用于机械臂和机器人跟踪动态目标的行为优化问题之中。 首先，分析了现有人工智能方法在解决实时决策问题上的不足之处，提出用于解决机器人系统进行实时决策的基于智能发育机制的场景驱动控制结构（MDCDCA）。不同于通常的基于模型的控制，该控制结构通过在知识库中检索当前的状态，找到最相近的知识作为决策结果。如果检索出的知识不满足规定的判据条件，将激活常规的优化算法，优化结果同时作为在线训练数据对知识库进行训练。这种控制结构不同于通常的离线训练方法，它可以从空白的知识库开始，在线发育知识。 其次，研究了MDCDCA控制结构的核心部分：智能发育模块。提出采用增强分级判别回归算法（IHDR）作为智能发育引擎。IHDR能够有效地实现从状态空间（Ｘ）到行为空间（Y）的映射。研究了IHDR涉及的相关技术，如：X和Y 空间的定义方法，知识的存储、积累、更新、以及检索等。根据其不足之处，介绍了改进的IHDR算法，以提高其实时性和健壮性。 第三，研究了将MDCDCA控制结构应用于7自由度机械臂动态目标跟踪，并将遗传算法作为MDCDCA的在线训练引擎。不同于最短时间规划，提出将施加在每个关节的力矩与其最大力矩的比值作为遗传进化过程的评价函数。因此，优化结果产生一个“最小疼痛感”轨迹。这类似于人手臂的运动轨迹，更加适合于像服务机器人这种与人相互协作的机器人系统。仿真结果表明了经过遗传算法训练的MDCDCA可以有效地完成轨迹规划任务，并满足系统的实时性要求。 最后，将在线遗传算法训练的MDCDCA方法应用于移动机器人在三维动态空间的路径规划，该问题单纯采用遗传算法在实时性上难以得到保证，仿真及对比结果表明所提出方法的实时特性，也表明所提出方法可用于具有类似特性的实时系统。

成就目标与心理健康的关系及相关因素的作用

The relationship between achievement goal and mental health, and the effect of the related factors Abstract Recently, adolescent’s mental health and depression problem have been widely concerned all over the world. Studies concerning it have been done from different perspectives. High school is the key period of adolescent’s physical and mental development, when they may have a high risk of mental conflicts, certain behavioral and emotional problems. Through questionnaire with high school students as subjects, this study examines achievement goal, self-concept, social comparison and the relationship between these factors and mental health. This study consists of three parts: 1) a preliminary exploration of the relationship between vocational school students’ achievement goal, self-concept and their mental health; 2) a revision of the simplified Chinese version for Dykman’ s goal Orientation Inventory; 3) a further discussion on the relationship between high school students’ achievement goal, self-concept, social comparison and depression. Major findings of this study are: 1. Dykman’s Goal Orientation Inventory is revised. This revised version, with high validity and reliability, conforms to Dykman’s theoretical assumptions and can be a convenient and effective measurement for achievement goal. 2. Self-concept and achievement goal of vocational school students can make a significant prediction of their mental health. Learning goal influences mental health through self-concept. 3. Achievement goal, self-concept and social comparison of high school students are closely related to their depression. 4. Self-concept and social comparison of high school students are the mediator variable between achievement goal and depression. The theoretical and practical implications of this study: 1. With the simplified Chinese version of Dykman’s Goal Orientation Inventory revised and used first in this study, its applicability is inspected and the conclusion drawn by Dykman on the prediction of achievement goal on depression is examined as well; on the basis of previous studies, a comprehensive and systematic study is done on how the factors work together to influence mental health(particularly depression). 2. This study enrichs the studies with high school students as subjects. It provides methods and basis for the theoretical studies and interference about high school students’ mental health education and depression problem.