Perception for mobile agents

移动代理的感知

• 批准号: 36744-2011
• 负责人: Jenkin, Michael
• 金额: $ 2.11万
• 依托单位: York 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=568775
• 关键词: Perception; mobile; agents

There are very few intelligent systems that do not exhibit mobility. But how do mobile systems manage to sense their environment while they move? My research program investigates fundamental issues related to how autonomous agents sense their environment as they move through it. This research program addresses this issue from three directions. First, it looks at how autonomous robots operating in a three dimensional world and equipped with passive sensors can build a 3D model of the environment as they move about. Whereas robots operating on the ground plane can use strong kinematic models to predict how sensor readings will change with vehicle motion, robots operating in three dimensions are typically unable to rely on such assumptions. This complicates somewhat the process of estimating robot motion and map construction. Second, this research program looks at how autonomous robots can estimate their motion and their orientation in their environment when they are unable to rely on a ground plane to define the direction of up and to help provide odometry estimates. Finally, this research program looks at how best to represent large scale 3D environmental structures obtained by autonomous systems operating the three dimensions. Solutions to these fundamental issues associated with robots operating in three dimensions in the real world will not only advance our understanding of autonomous systems but will also provide effective tools for a wide range of applications from understanding the decay of coral reefs to the development of effective technologies to support the use of robots in crime scene investigation.

很少有智能系统不具有移动性。但是，移动的系统如何在移动时感知环境呢？ 我的研究计划调查有关自主代理如何感觉他们的环境，因为他们通过它移动的基本问题。这个研究计划从三个方向解决这个问题。首先，它着眼于在三维世界中运行并配备被动传感器的自主机器人如何在移动时构建环境的3D模型。虽然在地平面上操作的机器人可以使用强运动学模型来预测传感器读数将如何随着车辆运动而变化，但在三维中操作的机器人通常无法依赖于这样的假设。这使得估计机器人运动和地图构建的过程变得复杂。其次，该研究计划着眼于自主机器人如何在无法依赖地平面来定义向上方向并帮助提供里程估计时，估计它们在环境中的运动和方向。最后，该研究计划着眼于如何最好地表示由自主系统操作的三维大规模三维环境结构。解决这些与机器人在真实的世界中的三维操作相关的基本问题，不仅将促进我们对自主系统的理解，而且还将为广泛的应用提供有效的工具，从了解珊瑚礁的腐烂到开发有效的技术，以支持在犯罪现场调查中使用机器人。