ITR/AP: Simulation of Machine-Medium Interaction in a Real-Time Virtual Environment

ghaboussi这个项目是一个多学科的产学研联合项目，旨在开发一个先进的虚拟现实（VR）环境，用于模拟土方设备与周围介质（如土壤）的相互作用。该项目将利用硬件、软件和信息技术的快速发展，为包括现实力反馈在内的机器-媒介交互开发一个实时虚拟环境。建议的发展将加强土方设备的设计，并缩短设计周期。它还将为VR在机器媒体交互中的应用开辟新的场所。该项目将极大地受益于卡特彼勒公司提供的杠杆资源。研究小组将开发一个基于原始神经网络（NN）的实时土壤介质模型，该模型可用于模拟土方设备操纵下土壤的响应。所提出的模型在机械上是准确的，并且是实时运行的。它将模拟土壤阻力和介质与土方设备之间的相互作用力。神经网络模型将使用使用离散元方法从非实时模拟中开发的数据集进行训练。土壤和土方设备响应的数据集也将从卡特彼勒公司在伊利诺伊州皮奥里亚试验场的全面现场测试中开发出来。研究小组将与国家超级计算应用中心（NCSA）合作，在洞穴自动虚拟环境（Cave）中实施新的神经网络土壤模型。在此虚拟环境下，将开发一个新的面向对象的车辆原型系统（VPS）。将建立一个力反馈环节，表示神经网络土壤模型提供给车辆动力学模型的介质阻力。实时虚拟环境将增强土方设备制造行业可用的设计验证工具，并减少新设备设计的昂贵现场试验次数。将为实时VR环境开发的思想和方法具有广泛的应用，例如模拟边坡破坏和雪崩，以及农业，采矿和制造业应用中的散装和粉末材料处理。所提出的虚拟现实环境可以潜在地用于其他行星的空间探索，其中需要实时排除与行星表面土壤介质相互作用的遥控车辆的故障。该项目汇集了来自伊利诺伊大学圣地亚哥分校土木与环境工程系的计算智能和土壤介质建模专家，伊利诺伊大学圣地亚哥分校NCSA的虚拟现实和计算机科学专家，以及卡特彼勒公司的土方设备设计和控制专家。团队成员之间的密切协调将确保学术界和工业界之间的思想交流

GhaboussiThis project is a joint multidisciplinary industry-academia research effort to develop an advanced virtual reality (VR) environment for modeling earthmoving equipment interaction with the surrounding medium such as soil. The project will take advantage of rapid developments in hardware, software and information technology to develop a real-time virtual environment for machine-medium interaction that includes realistic force-feed back. The proposed development will enhance the design of the earthmoving equipment and improve the design cycle. It will also open up new venues for application of VR for machine medium interaction. This project will greatly benefit from leveraged resources provided by Caterpillar, Inc.The research team will develop an original neural network (NN) based real-time soil medium model that can be used to simulate soil response due to manipulation by earthmoving equipment. The proposed model will be mechanistically accurate and run in real-time. It will simulate the soil resistance and the interactive forces between the medium and the earthmoving equipment. The NN model will be trained using data sets developed from non-real time simulations using the discrete element method. Data sets of soil and earthmoving equipment response will also be developed from full-scale field tests at Caterpillar, Inc. proving ground in Peoria, Illinois. The research team, in cooperation with the National Center for Supercomputing Applications (NCSA), will implement the new NN soil model in Cave Automated Virtual Environment (CAVE). A new object-oriented vehicle prototyping system (VPS) will be developed in this virtual environment. A force feedback link will be developed representing the medium resistance provided by the NN soil model to the vehicle dynamics model. The real-time virtual environment will enhance the design verification tools available to the earthmoving equipment manufacturing industry and reduce the number of costly field trials for new equipment designs. The ideas and methodology that will be developed for the real-time VR environment have a wide range of applications such as simulation of slope failures and avalanches as well as bulk and powder material handling in agricultural, mining and manufacturing applications. The proposed VR environment can potentially be used in space exploration of other planets where it is necessary to troubleshoot a remote controlled vehicle interacting with the soil medium on the planetary surface in real-time. This project brings together specialists in 1) computational intelligence and soil medium modeling from the Department of Civil and Environmental Engineering at UIUC, 2) virtual reality and computer science from NCSA at UIUC, and 3) earthmoving equipment design and control at Caterpillar, Inc. Close coordination among the team members will ensure a strong exchange of ideas between academia and industry.***