SitS NSF-UKRI: Rapid Deployment of Multi-Functional Modular Sensing Systems in the Soil

点击翻译按钮获取中文摘要

Soil is a global resource that supports most of our urban infrastructure, acts as a conduit for groundwater and is the dominant material in many of the world's geohazards. Understanding the in-situ state of soil (stress level, stiffness, strength, permeability) is essential to inform effective and efficient decisions about how humans should interact with soil deposits. Most current geotechnical measuring instruments involve vertical penetration of a probe into the soil to a shallow depth (up to a few hundred meters). Usually, the probe records only one type of data at a time (e.g. a displacement, a moisture content or a thermal gradient), in a very localized area. Consequently, the ground models used in decision-making rely on interpolation between relatively sporadic data points and consider relevant parameters (mechanical, hydraulic, thermal) separately. The Burrowing Robot with Integrated Sensing System (BRISS) builds on insight gained in designing Cone Penetration Test modifications and the more recent development of small prototype burrowing robots at the Georgia Institute of Technology (GT). The research objectives are to: (i) design, build and deploy a burrowing robotized sensor delivery system; (ii) sense mechanical and physical signals during the burrowing process and use machine-learning to adapt the burrowing process and the sensing strategy; (iii) interpret soil signals using particulate mechanics, tribology, large deformation continuum mechanics models and feature selection algorithms. The research group at GT, in collaboration with the research group at Imperial College London (ICL) in the UK, will collaborate to achieve the research objectives and to co-advise a cohort of graduate students and post-doctoral researchers. The BRISS will achieve a paradigm shift in soil exploration and site characterization. Both the sensor modules and the propulsion sections of the BRISS will be stackable, so that probes can be built up with different combinations of modules or the same modules but in different configurations. The BRISS will be minimally wired, thus the project findings will pave the way towards wireless, remotely controlled, multi-directional subsurface sensing. Such technologies will ultimately enable deep sediment characterization and extra-terrestrial exploration. The long-term deployment of multi-sensing probes could be used to detect variations of soil properties that are independent from localized probe stimuli, such as pH change consequent to mining activities or pore pressure change consequent to repeated droughts. The PIs will use GT and ICL institutional organizations to recruit students from under-represented minorities. They will engage with the ALERT Geomaterials network and GT Society of Women Engineers to attract female students. The lead-PI will participate in outreach activities for promoting the inclusion of the LGBTQ community in engineering and will facilitate Safe Space training for all the project team members.Challenges associated with obtaining undisturbed samples mean that probes that can measure these properties in-situ are incredibly useful. Informed by recent prototyping work at GT, the team will develop a novel multi-sensor system, BRISS, which will incorporate several major advances: (a) the use of soft robot and micro-controls to enable probes to navigate in any orientation in the subsurface; (b) the ability of these probes to self-propel through the soil using peristaltic motion; (c) the incorporation of multiple micro-sensors in these semi-autonomous probes; and (d) the leveraging of machine learning algorithms into the data analysis and soil model development. Recently developed experimental techniques will be used to refine and optimize the propulsion mechanism; these include novel textures, bio-inspired anchors, soil ablation mechanisms and self-lubrication processes. Innovative sensor systems will be designed and evaluated to optimize the set of measurements, with a particular focus on stress, stress wave velocity and acoustic emissions. Novel deep reinforcement machine learning algorithms will be used to refine the burrowing trajectory and adapt the frequency at which in situ measures are taken. Feature selection algorithms will be enhanced to handle large data sets for interpreting stress, pore pressure, geophysical and acoustic signals. This project will also provide fundamental understanding of the physical and mechanical response of dry and water-saturated sand during the penetration of the self-propelled BRISS. For the first time, multi-scale numerical models will decipher burrowing mechanics, by combining discrete element models that will include peristaltic boundary conditions, particle interaction and multi-scale tribological models that will shed light on the robot/soil interface rheology, and large-deformation finite element hydro-mechanical elasto-plastic models that will be applicable to predict soil behavior at larger scales.This project was awarded through the "Signals in the Soil (SitS)" opportunity, a collaborative solicitation that involves the United States Department of Agriculture National Institute of Food and Agriculture (USDA NIFA) and the following United Kingdom Research and Innovation (UKRI) research councils: 1) The Natural Environment Research Council (NERC), 2) the Biotechnology and Biological Sciences Research Council (BBSRC), 3) the Engineering and Physical Sciences Research Council (EPSRC), and the Science and Technology Facilities Council (STFC).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.