Rich interaction with the world requires extensive contact between robots and the objects in their environment. Most such contacts involve significant compliance between the contacting surfaces due to rubber pads or inflated grippers, soft objects to be manipulated, and soft surfaces for safe human-robot interaction. Accurate simulation of these contacts is critical for meaningful sim-to-real transfer. Compliant contact interactions generate contact surfaces of considerable extent, over which contact forces are distributed with varying pressure. Finite element methods can capture these effects but are too slow for most robotics applications. Consequently, in order to enable real-time simulation rates, most current simulation tools model contact as occurring between rigid bodies at a point or set of points using ad hoc methods to incorporate localized compliance. However, point contact is non-smooth, hard to extend to arbitrary geometry, and often introduces nonphysical artifacts. Moreover, point contact misses important area-dependent phenomena critical for robust manipulation, such as net contact moment and slip control. Pressure Field Contact (PFC) was recently introduced as a method for detailed modeling of contact interface regions at rates much faster than elasticity-theory models, while at the same time predicting essential trends and capturing rich contact behavior. PFC was designed to work with coarsely-meshed objects while preserving continuity to permit use with error-controlled integrators. Here we introduce a discrete approximation of PFC suitable for use with velocity-level time steppers that enables execution at realtime rates. We evaluate the accuracy and performance gains of our approach and demonstrate its effectiveness in simulation of relevant manipulation tasks. The method is available in open source as part of Drake's Hydroelastic Contact model.
Incorporating a dynamic kick engine that is both fast and effective is pivotal to be competitive in one of the world's biggest AI and robotics initiatives: RoboCup. Using the NAO robot as a testbed, we developed a dynamic kick engine that can generate a kick trajectory with an arbitrary direction without prior input or knowledge of the parameters of the kick. The trajectories are generated using cubic splines (two degree three polynomials with a via-point), cubic Hermite splines or sextics (one six degree polynomial). The trajectories are executed while the robot is dynamically balancing on one foot. Although a variety of kick engines have been implemented by others, there are only a few papers that demonstrate how kick engine parameters have been optimized to give an effective kick or even a kick that minimizes energy consumption and time. Parameters such as kick configuration, limits of the robot, or shape of the polynomial can be optimized. We propose an optimization framework based on the Webots simulator to optimize these parameters. Experiments on different joint interpolators for kick motions have been observed to compare results.
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.