Abstract: To solve the problem that children with autism have long duration and high medical expenses in the process of rehabilitation training,a human-computer interactive medical robot system for children with autism was designed and implemented.Through combining robot control technology with emotional computing,the multi-modal human-computer interaction emotional cognitive training method was designed.An emotion calculation model based on physiological signal sensitivity factor was proposed.A sensitivity factor based on multi-channel physiological signal fusion was used to modify the rehabilitation cognitive training process based on Radial Basis Function Neural Network (RBF_NN).Through the field test in a rehabilitation institution for autistic children in a community in Beijing,the feasibility and effectiveness of the system's rehabilitation treatment for assisted patients was verified.

Key words: robot, autism rehabilitation, human computer interaction, cognitive training, emotional calculation

摘要： 为解决孤独症儿童在康复训练过程中持续时间长、医疗费用高等问题,设计并实现了一种面向孤独症患儿的辅助康复医疗交互机器人系统。将机器人控制技术与情感计算相结合,设计了多模态人机交互的情感认知训练方式。提出了基于生理信号灵敏度因子的情感计算模型,根据径向基神经网络构建的一种基于多通道生理信号融合的灵敏度因子对康复认知训练过程进行修正。通过在北京市某社区孤独症儿童康复机构进行的实验测试,验证了该系统对辅助患者的康复治疗过程的可行性与有效性。

关键词: 机器人, 孤独症康复, 人机交互, 认知训练, 情感计算