Factors that influence the direction and timing of changes in travel path during a multiple obstacle avoidance task.

成功的导航需要人类和动物在充满必须避开的物体的环境中灵活机动。用于控制行动和安全导航到目标的行为变量是未知的。在混乱的环境中发展一般运动理论的一个挑战是确定个体如何协调和组织他们的行动。第一步是理解环境和个体是不断互动的，并受情景信息的指导。为了证明这一点，开发了转向动力学模型，以表明运动路径来自简单的导航规律。该模型假设目标导向行为可以用一些行为变量来描述，例如环境目标是轨迹趋同的吸引因素，而障碍是轨迹偏离的排斥因素。在过去的10年里，这个模型被广泛用于解释人类和非人类的导航。尽管这个模型对导航的某些方面提供了一个简单而优雅的描述，但当考虑在现实世界中导航时，它有几个明显的限制。该模型没有考虑个体的身体尺寸和障碍物的大小、动作能力，以及运动速度和方向的协调方式。为了解决这些限制，最近提出了一种完全不同的模型，即基于能力的模型。这种新方法表明，在选择动作和指导运动时，个体要考虑他或她的身体尺寸和运动能力。该模型的主要限制是，它要求个体已经学习并了解自己的身体尺寸和运动能力。这两个模型代表了不同的理论观点来解释用于指导运动的行为变量。理解个体或环境布局的内在属性对路径选择的影响一直没有答案，这一建议将为该领域以前的所有研究提供启示。我的基础研究项目的长期目标是确定两种模型中哪一种最能解释在现实世界中用于安全指导运动的行为变量。为了实现这一目标，将在未来五年内进行一系列研究，这些研究将操纵个人或环境的属性。这些研究将涉及现实世界的环境，参与者将沿着一条通往目标的道路行走，并避免与物理障碍相撞。在任何给定的试验中，障碍物将产生35到80厘米宽的间隙，并且在每个障碍物和间隙两侧的墙壁之间产生另外两条可能的路径。收集的数据将包括：1)关于路径选择的感性判断；2)凝视行为，反映感知信息的收集位置；3)反映路径选择的运动学。这项工作将决定这两个模型中哪一个能最好地解释在现实世界中用于安全指导运动的行为变量。这一提议背后的动机是将这种模型扩展到现实世界的情况下，在这种情况下，感知和行动必须与环境相协调。这将最终导致理解在混乱环境中控制运动的行为变量，这可以用来开发机器人安全地在不熟悉的环境中导航，或者在建筑物中正确规划出口路线。

Successful navigation requires both human and animals to maneouver through environments that are cluttered with objects that must be avoided. The behavioural variables used to control actions and safely navigate towards a goal are unknown. One challenge of developing a general theory of locomotion in cluttered environments is to determine how individuals coordinate and organize their actions. A starting step is to understand that the environment and the individual are in constant interaction and guided by episodic information. To demonstrate this, the Steering Dynamics Model was developed to show that locomotor paths emerge from simple navigation laws. The model assumed that goal-directed behaviours could be described by a few behavioural variables such that environmental goals served as attractors in which trajectories converged and obstacles served as repellors from which trajectories diverged. Over the past 10 years this model has been widely used to account for navigation in both humans and non-humans. Although this model offers a simple and elegant account of certain aspects of navigation, it has several significant limitations that are evident when one considers navigating in real-world situations. The model does not take into account individual’s physical size and size of the obstacles, action capabilities, or the manner in which locomotor speed and direction are coordinated. In order to address these limitations, a fundamentally different model, the Affordance-based Model, has recently been proposed. This new approach suggests that the individual takes into account his or her body dimensions and locomotor capabilities when choosing actions and guiding locomotion. The major limitation of this model is that it requires that an individual has learnt and has knowledge of his or her body dimensions and locomotor capabilities. These two models represent different theoretical ideas to explain the behavioural variables used to guide locomotion. Understanding the effects of intrinsic properties of an individual or environmental layout on path selection has long gone unanswered and this proposal will shed light on all of the previous research in the field. The long-term goal of my basic research program is to determine which of the two models best explains the behavioural variables used to safely guide locomotion in real-world situations. In order to meet this objective, a series of studies that manipulate either properties of the individual or those of the environment will be conducted over the next five years. These studies will involve real-world settings in which participants will walk along a path toward a goal and avoiding colliding with physical obstacles. The obstacles will create a gap between 35 and 80cm wide on any given trial as well as create two other possible pathways between each obstacle and the wall on either side of the gap. Data collected will include: 1) perceptual judgments about path selection; 2) gaze behaviours to reflect where perceptual information is being gather; and 3) kinematics to reflect path selections. This work will determine which of the two models best explains the behavioural variables used to safely guide locomotion in real-world situations. The motivation behind this proposal is to extend such models to real world situations where perception and action must coordinate with the environment. This will ultimately lead to understanding the behavioural variables that control locomotion in cluttered environments, which can be used to develop robots to safely navigate unfamiliar environments or properly plan exit routes in buildings.