Networking Strategies used by Bats to Improve Social Sonar

回声定位蝙蝠的生物声呐行为为如何优化自然和人工主动传感系统以满足快速变化的社会和环境条件提供了一个例子。蝙蝠和人工声纳、雷达和通信系统都面临着由相互干扰引起的一系列类似问题。解释蝙蝠在密集的群体中飞行或在嘈杂拥挤的栖息地中飞行时如何最大限度地减少对彼此声纳的干扰，可能有助于改善广泛的人工传感和无线通信系统。蝙蝠是高度社会化的，它们对同种干扰的特殊适应能力远远超过了目前人工系统中采用的策略。本提案将通过将计算机网络的最新进展应用于声纳/雷达理论来研究蝙蝠群体是如何做到这一点的。这些实验将描述蝙蝠在不同大小的群体和不同导航要求的任务中进行回声定位时，用来提高信号传递效率的行为策略。计算机模拟将用于确定最佳算法参数，并探索这些行为策略如何在更广泛的生物声纳条件下表现出来。通过破译蝙蝠可以合作共享相同声音空间的行为算法，该建议可能会导致更好和更有效的策略，以优化人工通信网络的使用，以支持日益信息化的经济。该提案还将为有兴趣探索生物系统作为人工系统灵感来源的STEM学生提供独特的接触，培训和教育机会。与当地科学教师合作，该提案创建了一个以探究为基础的双语（西班牙语/英语）教学网站，该网站将为五年级/六年级的科学班提供与活体动物远程互动的访问。在声纳和雷达、无线通信和计算机网络等人工系统中，当许多用户试图在一个共享信道上同时传输时，就会出现干扰问题。无线通信网络通过在高信道通信量期间采用限制用户传输速率的协议来尽量减少用户之间的干扰率，从而部分地减轻了这种情况。该提案验证了一个假设，即回声定位蝙蝠会根据类似的无线通信系统国际标准，调整其声呐脉冲发射率，以优化声呐在社交场合的性能。该提案结合了实验室对蝙蝠的行为研究和计算机模拟来解码蝙蝠的重传延迟算法。首先，训练有素的蝙蝠单独飞行，并以2-5为一组，通过传感器迷宫来量化蝙蝠的导航性能，并测量它们在社会环境中执行相同任务时是否系统地改变脉冲发射率。其次，在模拟其他蝙蝠通过迷宫的声音的人工声刺激下，让蝙蝠在迷宫中飞行，以测试蝙蝠是否可以通过降低自身脉冲发射率与种群密度成正比来实现声呐性能的净提高，从而通过减少脉冲发射获得信息流的净增加。将开发行为算法的计算模型来模拟行为，建立最佳参数并精确量化其在社会环境中对声纳性能的好处。蝙蝠的声纳算法将通过科学期刊提供，为该项目开发的基于matlab的生物声纳模拟工具将通过互联网上的多种来源免费提供。

The biosonar behavior of echolocating bats provides an example of how natural and artificial active sensing systems may be optimized to meet rapidly changing social and environmental conditions. Bats and artificial sonar, radar and communication systems all face a similar suite of problems arising from mutual interference. Explaining how bats minimize interfering with one other's sonar while flying in dense swarms or within noisy crowded roosts may help improve a wide range of artificial sensing and wireless communications systems. Bats are highly social and their exceptional resilience to jamming by conspecifics far exceeds the strategies currently employed in artificial systems. This proposal will investigate how groups of bats manage this by applying recent advances in computer networking to sonar/radar theory. The experiments outlined will characterize the behavioral strategies used by bats to improve their signaling efficiency when echolocating in groups of different sizes and across tasks of varying navigational demands. Computer simulations will be used to identify optimum algorithm parameters and explore how these behavioral strategies manifest across a wider range biosonar conditions. By deciphering the behavioral algorithms by which bats can cooperatively share the same acoustic space this proposal may lead to better and more efficient strategies for optimizing artificial communications network usage to support an increasingly information-based economy. This proposal will also offer unique exposure, training and educational opportunities for STEM students interested in exploring biological systems as a source of inspiration for artificial systems. In partnership with local science teachers, this proposal creates an inquiry-based dual language (Spanish/English) teaching website that will offer access for 5th/6th grade science classes to remotely interact with live animals. Interference problems arise in artificial systems such as sonar and radar, wireless communications and computer network when many users attempt to transmit simultaneously over a single shared channel. Wireless communications networks partially mitigate this by employing protocols constraining user transmission rates during periods of high channel traffic to minimize interference rates between users. This proposal tests the hypothesis that echolocating bats adapt their sonar pulse emission rates to optimize sonar performance in social situations following protocols similar and applicable to international standards in wireless communications systems. The proposal uses a combination of behavioral studies with bats in the lab and computer simulations to decode the bat's retransmission delay algorithm. First, trained bats are flown alone and in small groups of 2-5 through a sensor maze to quantify the bats' navigational performance and measure whether they systematically change pulse emission rates when performing the same task in a social context. Secondly, bats will be flown through the maze in the presence of artificial acoustic stimuli mimicking the sounds of other bats passing through the maze to test whether bats can realize a net improvement in sonar performance by decreasing their own pulse emission rates proportional to population density, counterintuitively reaping a net increase in information flow by emitting fewer pulses. Computational models of the behavioral algorithm will be developed to simulate the behavior, establish optimal parameters and precisely quantify its benefits to sonar performance in social contexts. The bat's sonar algorithms will be made available through scientific journals and Matlab-based biosonar simulation tools developed for the project will be made freely available via multiple sources on the Internet.