Abstract
Occlusions, restricted field of view and limited resolution all constrain a robot's ability to sense its environment from a single observation. In these cases, the robot first needs to actively query multiple observations and accumulate information before it can complete a task. In this paper, we cast this problem of active vision as active inference, which states that an intelligent agent maintains a generative model of its environment and acts in order to minimize its surprise, or expected free energy according to this model. We apply this to an object-reaching task for a 7-DOF robotic manipulator with an in-hand camera to scan the workspace. A novel generative model using deep neural networks is proposed that is able to fuse multiple views into an abstract representation and is trained from data by minimizing variational free energy. We validate our approach experimentally for a reaching task in simulation in which a robotic agent starts without any knowledge about its workspace. Each step, the next view pose is chosen by evaluating the expected free energy. We find that by minimizing the expected free energy, exploratory behavior emerges when the target object to reach is not in view, and the end effector is moved to the correct reach position once the target is located. Similar to an owl scavenging for prey, the robot naturally prefers higher ground for exploring, approaching its target once located.
Highlights
Despite recent advances in machine learning and robotics, robot manipulation is still an open problem, especially when working with or around people, in dynamic or cluttered environments (Billard and Kragic, 2019)
We develop a deep neural network architecture and training method to learn a generative model from pixel data consistent with the free energy principle, based on Generative Query Networks (GQN)
In this paper we investigated whether the active inference paradigm could be used for a robotic searching and reaching task
Summary
Despite recent advances in machine learning and robotics, robot manipulation is still an open problem, especially when working with or around people, in dynamic or cluttered environments (Billard and Kragic, 2019). A single sensory observation is not sufficient to capture the whole workspace, due to restricted field of view, limited sensor resolution or occlusions caused by clutter, human co-workers, or other objects. Humans on the other hand tackle this issue by actively sampling the world and integrating this information through saccadic eye movements (Srihasam and Bullock, 2008). They learn a repertoire of prior knowledge of typical shapes and objects, allowing them to imagine “what something would look like” from a different point of view. Recent work suggests that active vision and scene construction in which an agent uses its prior knowledge about the scene and the world can be cast as a form of active inference (Mirza et al, 2016; Conor et al, 2020), i.e., that actions are selected that minimize surprise
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
