Human activity learning for assistive robotics using a classifier ensemble

摘要: Assistive robots in ambient assisted living environments can be equipped with learning capabilities to effectively learn and execute human activities. This paper proposes a activity (HAL) system for application assistive robotics. An RGB-depth sensor is used acquire information of activities, set statistical, spatial temporal features encoding key aspects activities are extracted from the acquired Redundant removed relevant HAL model. ensemble three individual classifiers—support vector machines (SVMs), K-nearest neighbour random forest—is employed The performance proposed improved when compared other methods using single classifier. approach evaluated on experimental dataset created this work also benchmark dataset—the Cornell Activity Dataset (CAD-60). Experimental results show overall achieved by comparable state art has potential benefit applications reducing time spent