Collaborative Research: Increasing Capabilities of Heterogeneous Robot Teams through Mutually Beneficial Physical Interactions

建筑物(如房屋或公寓楼)倒塌后的城市搜救行动面临着复杂的布局，具有未知的危险，从火灾到带电的电线，再到可能发生的更多坍塌。这些未知因素和安全担忧导致搜索过程缓慢且经过深思熟虑，尽管时间是救援行动成功的最大预测因素之一。为幸存者确定家具或瓦砾造成“空口袋”的区域，以及为救援人员检查建筑物的安全性，是最费力和耗时的任务之一。具有不同能力的机器人组可以减少搜索区域，并提高人类操作员的安全性，大大有助于在这些环境中进行搜索。该奖项支持开发不同种类的机器人团队以帮助这些搜索和救援行动的研究。研究人员将为不同大小、形状和能力的机器人创造方法，让它们在共生团队中进行物理合作，并帮助彼此在坍塌的环境中移动。受植物启发的成长机器人将把能够填补缺口的小型网球大小的机器人运送到重要的搜索区域，而较小的机器人将帮助较大的成长机器人驾驶、感知和交流。从构建和研究这个共生机器人团队中学到的经验将有助于未来物理共生机器人的设计，并将作为开发新的搜救工具的起点。本研究探讨了共生异质机器人团队的发展，它利用团队成员之间的互惠互利的物理交互来共享现有能力并建立新的能力。研究人员将利用两种现有的机器人体系结构，即软增长机器人和微型机器人，在导航和探索非结构化环境方面具有互补的能力，并将研究如何设计这些机器人之间的物理交互来创造新的共生能力。虽然之前的工作主要集中在不同能力的机器人如何与同质团队相比提高团队绩效，但很少有工作证明不同能力的机器人在一起工作时可以如何增强彼此的能力。该项目直接研究如何设计这些共生行为，以使不同的机器人团队可以大于其各部分的总和，并有一个特定的目标，即创建团队来搜索灾难环境。特别是，这项工作将突出i)软增长机器人如何作为系绳和结构来扩展微型机器人的伸展范围，以及ii)具有不同有效载荷的微型机器人如何作为可变的驱动、传感和通信模块来扩展不断增长的机器人的运动范围和能力。研究人员将制作并研究一系列不同的潜在交互方式，同时确定和设计那些将最好地提高团队在模拟的非结构化区域中移动的能力的交互方式。总体而言，这项研究调查了高度多样化的软机器人和刚性机器人之间的互动，弥合了这些领域之间的差距，以探索如何利用每个领域的好处来创建更有效的机器人团队，以便更好地在非结构化世界中移动。这些结果将有助于机器人团队提高倒塌建筑中搜救队的安全性和成功率。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Urban Search and Rescue operations after a structure (e.g., a house or apartment building) collapse face complex layouts with unknown hazards, from fires to energized electrical wiring to the possibility of additional cave-ins. These unknowns and safety concerns lead to a slow and considered search process, even though time is one of the greatest predictors of success in rescue operations. Identifying areas where furniture or debris create “void pockets” for survivors and checking the safety of the structure for rescuers are some of the most laborious and time consuming tasks. Groups of robots with different abilities could reduce the areas to search and improve safety for human operators, significantly helping search efforts in these environments. This award supports research into the development of heterogeneous robot teams to aid in these search and rescue operations. The researchers will create methods for robots of different sizes, shapes, and abilities to physically work together in symbiotic teams and help each other move around collapsed environments. Small tennis-ball-sized robots that can fit through gaps will be delivered to important search areas by plant-inspired growing robots, and the smaller robot will help the larger growing robot steer, sense, and communicate. The lessons learned from building and studying this symbiotic robot team will help future designs for physically symbiotic robots and will serve as a starting point for developing new search and rescue tools. This research investigates the development of symbiotic heterogeneous robot teams which use mutually beneficial physical interactions between team members to share existing capabilities and build new ones. The researchers will leverage two existing robot architectures, soft growing robots and microrobots, with complementary capabilities in the area of navigation and exploration of unstructured environments, and will study how physical interactions between these robots may be designed to create new symbiotic capabilities. While prior work has focused on how robots of differing capabilities can improve team performance compared to homogeneous teams, little work has demonstrated how heterogeneous robots may augment each other's capabilities when they function together. This project directly investigates how these symbiotic behaviors may be designed so that heterogeneous robot teams can be greater than the sum of their parts, with a specific target of creating teams to search disaster environments. In particular, the work will highlight how i) soft growing robots can serve as tethers and structures to extend the reach of microrobots and ii) how microrobots with different payloads can act as changeable actuation, sensing, and communication modules for expanding a growing robot’s range of motion and capabilities. The researchers will prototype and study a diverse range of potential interaction styles while identifying and designing those interactions that will best improve the ability of the team to move in simulated unstructured areas. Overall, the research investigates interactions between highly diverse soft and rigid robots, bridging the gaps between these fields to explore how benefits from each can be leveraged to create more effective robotic teams that can better move around the unstructured world. These results will build towards robot teams which can improve safety and success of search and rescue teams in collapsed buildings.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.