Deep reinforcement learning has been successfully used in many dynamic decision making domains, especially those with very large state spaces. However, it is also well-known that deep reinforcement learning can be very slow and resource intensive. The resulting system is often brittle and difficult to explain. In this paper, we attempt to address some of these problems by proposing a framework of Ruleinterposing Learning (RIL) that embeds high level rules into the deep reinforcement learning. With some good rules, this framework not only can accelerate the learning process, but also keep it away from catastrophic explorations, thus making the system relatively stable even during the very early stage of training. Moreover, given the rules are high level and easy to interpret, they can be easily maintained, updated and shared with other similar tasks.
Undesirable chatter is one of the key problems that restrict the improvement of robot milling quality and efficiency. The prediction of chatter stability, which is used to guide the selection of process parameters, is an effective method to avoid chatter in robot milling. Due to the weak stiffness of the robot, deformation caused by milling forces becomes an unavoidable problem, which will change the tool–workpiece contact area and affect the stability prediction. However, it is often simplified and neglected. In this paper, a multipoint contact dynamic model of robot milling is established, which considers the influence of force-induced deformation on the regenerative effect and process damping. The tool–workpiece contact area is discretized into a finite number of nodes along the axial direction so that the force and deformation at each node can be calculated separately. The different contact forms of the tool–workpiece under different process parameters are discussed in different cases, and the interaction process between cutting force and force-induced deformation is analyzed in detail. An iterative strategy is used to calculate the deformation of each node and the result of the tool–workpiece contact boundary. Finally, chatter stability of robot milling is predicted by a fully discrete method. Robot milling experiments were carried out to verify the predicted results. The results show that force-induced deformation is an important factor improving the stability prediction accuracy of robot milling, and a more accurate prediction result can be obtained by simultaneously considering force-induced deformation and process damping.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.