RI: Small: Robotic Search of Transient Objects

在物理空间中搜索物体是机器人最常见的任务之一。瞬时目标是指除非满足瞬时感测和/或信令条件，否则无法识别的一类目标。瞬变特性通常由目标属性、隐私问题、环境约束和/或感测限制引入。例如，在飞机坠毁后在海底搜索黑匣子是一项典型的瞬时目标搜索(TTS)任务，因为搜索依赖于黑匣子发出的瞬时无线电/声纳信号。该项目为TTS问题奠定了理论基础，这些问题将影响一大组现实世界的应用程序。由于TTS问题的随机性，TTS问题具有挑战性，因为瞬变特性通常与诸如感知范围限制、各种覆盖函数、受限的移动性、信号对应、有限的搜索者数量和广阔的搜索区域等因素相耦合。将进行广泛的建模和分析，以了解搜索过程中的各种因素。因此，将开发一个新的计算框架，包括整合资源以提高TTS性能的模型、度量和算法，并将重点放在三个方面：解决关键性能问题的搜索时间分析，关注唯一信号对应和协作感知问题的多目标搜索，以及包括集中式和分散式两种方法的机器人协调规划。该项目还为本科生和研究生开发了新的课程材料，以更好地了解机器人搜索应用程序中的算法和稳健的智能。

Searching for objects in physical space is one of the most common tasks for robots. Transient targets refer to a class of objects which are not identifiable unless momentary sensing and/or signaling conditions are satisfied. The transient property is often introduced by target attributes, privacy concerns, environment constraints, and/or sensing limitations. For example, searching for a black box on the ocean floor after an airplane crash is a typical transient target search (TTS) task because the search relies on the transient radio/sonar signals emitted from the black box. This project develops the theoretical foundations for TTS problems that will impact a large group of real world applications. Due to their stochastic nature, TTS problems are challenging because the transient property is often coupled with factors such as sensing range limits, various coverage functions, constrained mobility, signal correspondence, limited number of searchers, and a vast searching region. Extensive modeling and analysis will be performed to understand the various factors in the searching process. As the result, a new computation framework consisting of models, metrics, and algorithms that integrate the resource to improve TTS performance will be developed with three focuses: searching time analysis that addresses the critical performance issue, multi-target search that focuses on unique signal correspondence and collaborative sensing issues, and planning for coordination of robots which includes both centralized and decentralized approaches. This project also develops new course materials for undergrad and graduate students to better understand algorithms and robust intelligence in robotic searching applications.