Blended Intelligence for Safe and Efficient Nuclear Sort & Segmentation

威立雅核解决方案的财团汇集了来自领先的核能、农业技术领域的成熟技术和深厚专业知识。，空间和人工智能从业者应对核废料分类和细分的挑战。该项目将提供一个能够在非活动环境中验证关键系统性能参数的演示器。我们的系统将使用一个类似于人类手臂的远程操纵器，它与先进的特征和跟踪技术一起工作，以识别、分类、拾取、转移、跟踪和包装核废料。机械手手臂的操作可以在计算机智能控制下自主操作，或者对于更复杂的抓取和纠结的废弃物，需要有专业知识的熟练操作人员来接管任务。通过在人类和计算机智能之间共享高度异构废物的分类和分割任务，我们的目标是在当前过程的安全性和效率方面提供一个步骤改变，复杂任务只需要熟练操作员的有限干预。我们的模块化方法侧重于建立正确的智能和决策框架，以决定如何管理和分类废物，以及正确的废物处理能力，以分离废物并将其送到需要的地方。用于表征的输入传感器，以及机械手平台的配置，设计得很容易改变，以适应被分类的废物，使该方法可在废物流和分类位置之间转移。该技术提高了操作员的自主操作能力，同时将他们从危险环境中移除。效率取决于每个操作人员部署的机械手数量。重要的是，这为废物处理从业者提供了灵活性，可以根据流程确定人工/自主能力的正确组合，并随着系统和操作员的学习而随着时间的推移而改进。领导该项目的是威立雅核能解决方案英国公司（VNS），核机器人专家和DEXTERTM的开发商。VNS将得到林肯大学的支持，该大学提供自动化和力制导技术，学院提供用于废物跟踪，包装和物理化学表征的人工智能，Createc放射检测设备专家以及莫特麦克唐纳废物管理顾问。

Veolia Nuclear Solutions' consortium brings together proven technologies and deep domain expertise from leading nuclear, agritech., space and artificial intelligence practitioners to meet the challenge of nuclear waste sort and segmentation. This project will deliver an demonstrator capable of validating key system performance parameters in an inactive environment.Our system will use a tele-manipulator, analogous to a human's arms, that works together with advanced characterisation and tracking technologies to identify, classify, pick up, transfer, trace and package nuclear waste . Operation of the manipulator arms can be under the control of computer intelligence and operate autonomously, or for more complex grabs and tangled wastes, call on the expertise of a skilled operator to take over the task. By sharing the task of the sorting and segmentation of highly heterogeneous wastes between human and computer intelligence, we aim to offer a step change in the safety and efficiency of the current process, only limited intervention by skilled operators will be needed for complex tasks.Our modular approach is focused around building the right intelligence and decision making frameworks to decide how to manage and classify the waste, and the right waste manipulation capability to separate the waste and get it where it needs to go. Input sensors for characterisation, and configuration of the manipulator platform, are designed to be easily changed to suit the waste being sorted making the approach transferable across waste streams and sorting locations.The technology enhances operator capability with autonomous operation whilst removing them from the hazardous environment. Efficiency scales as the number of manipulator units deployed per operator. Importantly this builds in flexibility for waste practitioners to determine the right mix of human/autonomous capability by stream and improve with time as the system and operators learn.Leading the project is Veolia Nuclear Solutions UK (VNS), an expert in nuclear robotics and developer of DEXTERTM. VNS will be supported by the University of Lincoln providing automation and force guidance technology, Faculty providing the artificial intelligence for waste tracking, packing and physico-chemical characterisation, Createc a specialist in radiological detection equipment, and Mott Macdonald waste management consultants.