arXiv Science⌕ Search

arXiv subjects

Paul Montgomery

Publications and source records attributed to Paul Montgomery.

3 recordsLinked to original sources

Heavy Metal Acrobatics: Learning Dynamic Control for Skid-Steer Loaders

Automation of heavy machinery offers the opportunity to improve productivity in mechanized industries such as construction, mining, and material handling. Yet, automation of such machines is challenged by their complex dynamics, in part because of hydraulic actuation that involves fluids, flexible elements, and nonlinear delays. A concomitant challenge is the size and weight of these machines, which can make experiments on the physical systems exceedingly dangerous. Here, we show that \ac{RL} in simulation, followed by transfer of the learned policies from simulation to reality, can successfully automate acrobatic maneuvers that push a heavy construction machine to the limits of its actuation and have only previously been accomplished by a small number of elite human operators in official world-record performances. We trained policies that perform autonomous acrobatic maneuvers on a John Deere 332G, an operational 5-tonne skid-steer loader. The policies can autonomously swing the machine up onto two wheels and, while balancing on two wheels, drive over rough terrain, trace out specified trajectories, spin around, and swing the bucket back and forth before landing it safely back onto four wheels. This work demonstrates that \ac{RL} and sim-to-real transfer enable dynamic autonomous control of heavy machinery, directly contributing to greater automation in heavy equipment sectors. Videos: https://sites.google.com/leggedrobotics.com/heavy-metal-acrobatics This is the author's version of the work. It is posted here by permission of the AAAS for personal use, not for redistribution. The definitive version was published in Science Robotics on 11 November 2026, DOI: https://doi.org/10.1126/scirobotics.aef0894

cs.RO↗

Super-resolution imaging within reach

Although several optical techniques have been recently developed in order to overcome the resolution limit in microscopy, the imaging of sub-wavelength features is still a real challenge. In practise, super-resolution techniques remain difficult to build or are photo-toxic for the biological samples. However, microsphere-assisted microscopy has recently made super-resolution imaging accessible to scientists (e.g. optical metrologists, engineers and biologists). This paper presents an easy-to-implement optical setup to perform full-field and contactless super-resolution measurements of nanostructured media or biological elements. For this purpose, a classical microscope was enhanced by introducing a transparent microsphere. We show that this rather simple approach makes it possible to achieve a lateral resolution of 200 nm in air, i.e. the visualization of feature sizes of 100 nm.

physics.ed-ph↗

Fourier optics: basic concepts

Based on diffraction theory and the propagation of the light, Fourier optics is a powerful tool allowing the estimation of a visible-range imaging system to transfer the spatial frequency components of an object. The analyses of the imaging systems can thus be performed and the the performance retrieved. For a better understanding of the optical study, we present a short introduction of the Fourier optics and we review the mathematical treatment depending on the illumination conditions of the imaging system. Furthermore, resolution criteria based on Fourier optics are detailed. Also, the impact of aberrations on the imaging quality are discussed.

physics.gen-ph↗