Mechanism Design and Control of Bio-inspired Inspection Robots

在过去的十年中，工业对使用机器人进行机械结构检测的需求越来越大。然而，一些复杂机械结构的小、受限区域给使用工业机器人进行可靠性、效率和可操作性的检测带来了挑战。仿生机器人模仿一些动物的行为来维持一个动态平衡的系统，为这些问题提供了解决方案，因此成为最近的研究趋势。这项研究计划旨在解决目前使用机器人进行检查的障碍：机动性差，效率低和工作区域有限。一个典型的例子是水下海洋结构或具有密闭区域的机体结构的检测。由于具有障碍物和孔洞的结构的进出受限和复杂性，目前的结构检测仍然高度依赖于人工操作。本研究计划提出了一种多足仿生机器人的设计，用于检测具有受限区域的复杂机械结构并收集数据。检测机器人被期望能够深入到被检测结构中，并具有高机动性和避障能力。本文的研究包括机构设计、传感系统和控制方法。创新的传动系统，基于我们之前在麦吉尔大学智能机器中心为移动机器人设计的传动系统，提出了实现机器人每条腿三个移动的方案。通过这样做，所提出的检测机器人在提高机动性方面将比目前市场上的多足机器人具有显著的优势，可以在崎岖地形上进行运动并避开障碍物。该微尺度传感系统采用最先进的MEMS制造技术，有望产生机器人进行三维运动的姿态、位置、方向及其扭转点速度和角速度的准确信息。本研究也将提出一种基于神经网络和遗传算法的智能控制方法。拟议的机器人设计将与水下航行器集成，以实现海洋工程目的的完全自主数据收集。通过与不同研究小组的密切合作，拟议研究的理论发展预计将服务于广泛的科学和工业部门，包括水下导航和数据收集，石油管道检查和机身结构检查。预计该研究计划将产生具有重大影响的成果丰硕的出版物，并培养3名博士和2名硕士学生作为该计划的一部分。

The past decade has witnessed the increasing demands from industry for using robots in the inspection of mechanical structures. However, small, confined areas of some complex mechanical structures brings challenges for using industrial robots to perform inspections considering reliability, efficiency and manoeuvrability. Bio-inspired robots, mimicking the behaviors of some animals to maintain a dynamically balanced system, provide solutions to some of these issues, thus, becoming a research trend recently. This research proposal is targeting to address the current hurdles of using robots for inspection: poor manoeuvrability, low efficiency and limited working areas. A typical example is the inspection of underwater marine structures or airframe structures with confined areas. Due to the limited access and complexity of structures with obstacles and holes, the structure inspection is still highly dependent on human operators nowadays. This research program proposes the design of a multilegged bio-inspired robot for the purpose of inspection of complex mechanical structures with confined areas and data collection. The inspection robot is expected to move deeply into the structure under inspection with high manoeuvrability and obstacle avoidance capability. The proposed research is composed of mechanism design, sensing system and control methodology. Innovative transmission system, based on our previous transmission system design for mobile robots at the Centre for Intelligent Machines at McGill University, is proposed to realize mobility of three for each leg of the robot. By doing this, the proposed inspection robot would have significant advantages over the current multilegged robots on the market in improving manoeuvrability to carry out locomotion over rough terrain and avoid obstacles. The micro-scale sensing system, using the state-of-the-art MEMS fabrication technology, is expected to yield the accurate information of pose-position and orientation-and its twist-point velocity and angular velocity of robot undergoing 3D motions. This research will also propose developing an intelligent control methodology based on neural networks and genetic algorithms. The proposed robot design will be integrated with underwater vehicles, to realize fully autonomous data collection for ocean engineering purposes. Through close collaboration with diverse research groups, the theoretical developments of the proposed research are expected to serve a wide range of scientific and industrial sectors, including underwater navigation and data collection, oil pipeline inspection and airframe structure inspection. It is expected that the proposed research program will result in fruitful publications with significant impact and training 3 PhD and 2 Master's students as part of this program.