项目名称: 普适环境下基于三维肢体动作理解的助老机器人交互技术研究
项目编号: No.61273339
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 李瑞峰
作者单位: 哈尔滨工业大学
项目金额: 79万元
中文摘要: 普适环境的肢体动作理解是助老服务机器人当前研究的热点问题,同时也是助老机器人走向实际应用必须解决的难点问题,该问题解决对助老机器人认知和老人陪护问题具有重要意义。本项目拟采用通过老人容易掌握的肢体形态方式建立新机制,解决在普适环境下助老机器人的人机自然交互问题。在助老机器人平台下,通过基于机器人本体获取的深度和图像信息的融合,实现在普适环境对人物目标区域的识别和准确提取,构建实时的三维人体姿态完成对肢体动作的识别,并结合行为心理学知识库对肢体动作进行分析,针对老人行为动作的理解建立人机交互策略。该方案对室内光线变化、复杂背景、局部遮挡和动态因素干扰具有鲁棒性特点,在人机交互中具有稳定性和实时性特点,克服了传统方法受环境和用户变化影响的缺点。该项目研究不仅提高助老机器人的人机交互能力和智能化程度,为助老机器人的人机交互方向提供理论支持,而且对助老机器人的实际应用具有重
中文关键词: 助老服务机器人;人体姿态识别;肢体动作理解;人机交互;
英文摘要: To resolve the problems of eldercare robot encounted in the universal environment, a novel method will be developed in this scheme. Novel modeling will be developed to capture human actions which is sutiable for elder to solve the natural human-robot-interaction(HCI) problems in this scheme. Based on the eldercare robot platform, depth image and color image are fused to detect the target from the background, and to estimate the 3D human pose for realtime action recognition. To analysis the motion of elder using the psychology database, HCI strategy will be bulit based on action perception. This proposed scheme exhibits robust, stable and real-time performance which are well suited for the the unpredicted changes of indoor light, complex background, partial occlusion and other facoctos, thus overcoming the shortage in traditional method affected by the changes of environment and user. This scheme not only promotes the HCI and Artificial Intelligence technology for the robot theoretical significance, but also has important realistic meaning to push eldercare robot in application.
英文关键词: assistant robots for elderly;human gesture recognition;body action perception;human-robot interaction;