Spatial learning strategies, visual information acquisition and efficiency of local navigation

一些昆虫，如蜜蜂或蚂蚁，表现出非凡的能力来学习它们食物地点的空间位置并返回这些地方。在蜜蜂中，与目标附近导航相关的视觉信息通过结构化学习飞行进行积极组织。尽管很明显，学习飞行在建立目标位置的视觉表征方面发挥了重要作用，但在很大程度上仍不清楚(1)这种行为的哪些组成部分与寻找目标有关，(2)蜜蜂使用什么机制从复杂的时空视网膜输入中提取相关信息，(3)蜜蜂如何利用这些信息来建立空间表征和控制行为，以及(4)这些机制如何应对环境中不可预测的变化。回答这些根本问题将是拟议项目的主要目标。新的方法将被用来将高速摄像机监控的行为序列分离成不同的组成部分，并重建蜜蜂在学习和寻找目标的飞行过程中所看到的情况。行为分析将在下一个项目阶段通过在3D空间中移动人工视觉系统的龙门式机器人建模和实施局部导航的生物学原理来补充。因此，该项目虽然主要涉及基础研究，但从长远来看，很可能与人工自主导航系统相关。

Some insects, such as bees or ants, demonstrate an extraordinary ability to learn the spatial location of their food sites and to return to these places. In bees the visual information relevant for navigation in the vicinity of the goal is actively organized by structured learning flights. Although it is clear that learning flights play an important role in establishing a visual representation of the goal location, it is still largely unknown (1) which components of this behavior are relevant to finding the goal, (2) what mechanisms bees employ to extract the relevant information from the complex spatio-temporal retinal input, (3) how bees use this information to establish spatial representations and to control behavior and (4) how these mechanisms can cope with unpredictable changes in the environment. To answer these fundamental questions will be the main objective of the proposed project. Novel approaches will be employed to separate behavioral sequences monitored with high-speed cameras into distinctive components and to reconstruct what the bee has seen during her learning and goal-seeking flights. The behavioral analysis will be complemented in a next project phase by modeling and implementing biological principles of local navigation in a gantry robot moving an artificial vision system in 3D space. Hence, this project, though mainly concerned with basic research, is, in the long run, likely to be relevant for artificial autonomous navigation systems.