Project Details
Projekt
Cognitive modeling of the predictive active self in situated action (COMPAS)
Subject Area
General, Cognitive and Mathematical Psychology
Term
since 2018
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 402792175
Autonomous embodied agents in an uncertain real-world environment need to continuously adapt their situated action. One important mechanism for this is a sense of control (SoC) , i.e. the assessment of how one is in control of one's own actions and their effects. This project investigates how the SoC arises from and affects the cognitive and sensorimotor processes underlying the control of situated actions. To that end, we combine experiments with human participants with computational cognitive modeling and simulation-based studies. In the previous project phase we have (1) implemented the first cognitive architecture for prediction-based action control by integrating an established model of cognitive processes (ACT-R) with a probabilistic model of predictive sensorimotor processes, (2) modeled the sensorimotor and cognitive components of SoC as well as their interplay in the top-down/bottom-up processes underlying prediction-based action control, (3) designed and applied a task scenario for experimental data collection and model testing through simulation. In the present phase we will build on these outcomes and pursue three threads of research. First, we will evaluate our conceptual, methodological and modeling approach by studying whether we are able to predict the behavior and SoC of human participants from situational aspects. For this we will explore and utilise a range of measure such as temporal binding, eye-tracking and EEG. Second, we will investigate the extent to which the active self is imparted to task-specific representations and thus influences the cognitive processes that control how concurrent tasks are managed. Finally, we will study how SoC arises in situated action control with multiple sensory input modalities (visual, acoustic, haptic). The work program consists of five experimental studies along with work on extending the conceptual-theoretical model of SoC and its technical implementation in terms of a concrete computational cognitive model that can be run in simulations and will be systematically tested against the experimental data. Concrete collaborations with other subprojects of the SPP are planned to facilitate a common understanding and methodology to study the active self.
DFG Programme
Priority Programmes
Subproject of
SPP 2134:
The active self
