Kitchen Task Assessment Dataset for Measuring Errors due to Cognitive Impairments

The dataset contains different types of sensor data: acceleration data from different objects and from body-worn sensors as well as annotation. The dataset consists of 12 normal runs and 12 erroneous runs, where the participants simulated typical errors due to dementia. The annotation consists of both action annotation in the form “action_object_object” as well es annotation about the object being manipulated and the hand that is manipulating it.