Autonomous Exploration of Challenging Environments

挑战环境的自主探索

• 批准号: 356377-2013
• 负责人: Rekleitis, Ioannis
• 金额: $ 1.09万
• 依托单位: McGill University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=572757
• 关键词: Autonomous; Exploration; Challenging; Environments

The problem of exploring an unknown environment is one where robotic research pushes the boundaries of knowledge and technology. Coral reef monitoring with underwater vehicles, planetary exploration, aerial surveys for environmental monitoring or for search and rescue all share some similar characteristics despite the diversity of the domains. In all the above domains, a series of common problems need to be addressed. First, the robot should be able to localize itself by fusing information from proprioceptive sensors, and exteroceptive sensors. Algorithms for estimating the position and orientation (pose) of a mobile robot are considered significant enablers for robot autonomy, and thus, a large spectrum of robotics research has focused on them. The second fundamental problem in autonomous exploration is planning the trajectory of the autonomous vehicle in order to ensure accurate estimation and modeling of the environment while at the same time satisfying the criteria of completeness and efficiency. In recent work I have developed an algorithm for the complete visual coverage of known environment using an unmanned aerial vehicle in an efficient manner. In addition my work on balancing exploitation and exploration, that is the trade-offs between efficiency and accuracy, which was developed for indoor environments using a graph representation, would be extended for the underwater domain. The above components would enable robots to accurate collect sensory information from the underwater domain in a systematic and efficient manner. Underwater environments are particularly challenging due to limited visibility, locally self-similar structures, and the unpredictability of motions due to surge and currents. The final part of the proposed research would be to develop off-line bundle-adjustment based algorithms for the accurate reconstruction of environmental models combining position, intensity, and color information for used by scientist in the fields of marine biology and planetary exploration [J3]. In particular, autonomous operations over coral reefs would benefit greatly from accurate localization algorithms.

探索未知环境的问题是机器人研究突破知识和技术界限的问题。利用水下航行器监测珊瑚礁、行星探索、用于环境监测或搜救的航空勘测都有一些相似的特点，尽管领域各不相同。在上述所有领域，都需要解决一系列共性问题。首先，机器人应该能够通过融合来自本体感知传感器和外部感知传感器的信息来定位自己。移动机器人的位置和姿态估计算法被认为是机器人自主的重要因素，因此，广泛的机器人学研究都集中在这些算法上。自主探索的第二个基本问题是规划自主车辆的轨迹，以确保在满足完整性和效率标准的同时对环境进行准确的估计和建模。在最近的工作中，我开发了一种算法，可以高效地使用无人机对已知环境进行完全可视覆盖。此外，我在开发和探索之间的平衡工作，也就是效率和精度之间的权衡，是使用图形表示为室内环境开发的，将扩展到水下领域。 上述部件将使机器人能够以系统和有效的方式从水下领域准确地收集感官信息。水下环境由于能见度有限、局部自相似结构以及由于浪涌和洋流而产生的运动的不可预测性而特别具有挑战性。拟议研究的最后部分将是开发基于离线捆绑平差的算法，以准确重建结合位置、强度和颜色信息的环境模型，供科学家在海洋生物学和行星探测领域使用[J3]。特别是，珊瑚礁上的自主作业将从准确的定位算法中受益匪浅。