Search

The design of gaits for robot locomotion can be a daunting process which requires significant expert knowledge and engineering. This process is even more challenging for robots that do not have an accurate physical model, such as compliant or micro-scale robots. Data-driven gait optimization provides an automated alternative to analytical gait design. In this paper, we propose a novel approach to efficiently learn a wide range of locomotion tasks with walking robots. This approach formalizes locomotion as a contextual policy search task to collect data, and subsequently uses that data to learn multi-objective locomotion primitives that can be used for planning. As a proof-of-concept we consider a simulated hexapod modeled after a recently developed microrobot, and we thoroughly evaluate the performance of this microrobot on different tasks and gaits. Our results validate the proposed controller and learning scheme on single and multi-objective locomotion tasks. Moreover, the experimental simulations show that without any prior knowledge about the robot used (e.g., dynamics model), our approach is capable of learning locomotion primitives within 250 trials and subsequently using them to successfully navigate through a maze., 8 pages. Accepted at RAL+ICRA2018

We present the instrument design and first light observations of KAPAO, a natural guide star adaptive optics (AO) system for the Pomona College Table Mountain Observatory (TMO) 1-meter telescope. The KAPAO system has dual science channels with visible and near-infrared cameras, a Shack-Hartmann wavefront sensor, and a commercially available 140-actuator MEMS deformable mirror. The pupil relays are two pairs of custom off-axis parabolas and the control system is based on a version of the Robo-AO control software. The AO system and telescope are remotely operable, and KAPAO is designed to share the Cassegrain focus with the existing TMO polarimeter. We discuss the extensive integration of undergraduate students in the program including the multiple senior theses/capstones and summer assistantships amongst our partner institutions. This material is based upon work supported by the National Science Foundation under Grant No. 0960343.

We describe KAPAO, our project to develop and deploy a low-cost, remote-access, natural guide star adaptive optics (AO) system for the Pomona College Table Mountain Observatory (TMO) 1-meter telescope. We use a commercially available 140-actuator BMC MEMS deformable mirror and a version of the Robo-AO control software developed by Caltech and IUCAA. We have structured our development around the rapid building and testing of a prototype system, KAPAO-Alpha, while simultaneously designing our more capable final system, KAPAO-Prime. The main differences between these systems are the prototype's reliance on off-the-shelf optics and a single visible-light science camera versus the final design's improved throughput and capabilities due to the use of custom optics and dual-band, visible and near-infrared imaging. In this paper, we present the instrument design and on-sky closed-loop testing of KAPAO-Alpha as well as our plans for KAPAO-Prime. The primarily undergraduate-education nature of our partner institutions, both public (Sonoma State University) and private (Pomona and Harvey Mudd Colleges), has enabled us to engage physics, astronomy, and engineering undergraduates in all phases of this project. This material is based upon work supported by the National Science Foundation under Grant No. 0960343., 10 pages and 11 figures