Cognitive Mapping and Planning for Visual Navigation
Cognitive Mapping and Planning for Visual Navigation
复制标题
DOI：
10.1007/s11263-019-01236-7
复制
发表时间：
2017-02
影响因子：
19.5
通讯作者：
Saurabh Gupta;Varun Tolani;James Davidson;S. Levine;R. Sukthankar;J. Malik
中科院分区：
文献类型：
--
作者：
Saurabh Gupta;Varun Tolani;James Davidson;S. Levine;R. Sukthankar;J. Malik
我们介绍了一种用于新环境下导航的神经结构。我们提出的架构学习从第一人称视角绘制地图，并计划在环境中实现目标的一系列行动。认知映射器和规划器（CMP）基于两个关键思想：a)映射和规划的统一联合架构，这样映射就由规划器的需求驱动；b)空间记忆，具有在对世界的不完整观察集的情况下进行规划的能力。CMP构建了自上而下的世界信念图，并应用可微神经网络计划器在每个时间步产生下一个动作。对世界的累积信念使智能体能够跟踪环境中访问过的区域。我们的实验表明，CMP优于反应策略和标准的基于内存的架构，并且在新环境中表现良好。此外，我们表明CMP还可以实现语义指定的目标，例如“到椅子上”。
We introduce a neural architecture for navigation in novel environments. Our proposed architecture learns to map from first-person views and plans a sequence of actions towards goals in the environment. The Cognitive Mapper and Planner (CMP) is based on two key ideas: a) a unified joint architecture for mapping and planning, such that the mapping is driven by the needs of the planner, and b) a spatial memory with the ability to plan given an incomplete set of observations about the world. CMP constructs a top-down belief map of the world and applies a differentiable neural net planner to produce the next action at each time step. The accumulated belief of the world enables the agent to track visited regions of the environment. Our experiments demonstrate that CMP outperforms both reactive strategies and standard memory-based architectures and performs well in novel environments. Furthermore, we show that CMP can also achieve semantically specified goals, such as" go to a chair".