Mapping Navigation Instructions to Continuous Control Actions with Position-Visitation Prediction
Mapping Navigation Instructions to Continuous Control Actions with Position-Visitation Prediction
复制标题
DOI：
--
复制
发表时间：
2018-10
期刊：
影响因子：
--
通讯作者：
Valts Blukis;Dipendra Misra;Ross A. Knepper;Yoav Artzi
中科院分区：
文献类型：
--
作者：
Valts Blukis;Dipendra Misra;Ross A. Knepper;Yoav Artzi
我们提出了一种绘制自然语言指令和原始观察的方法，以连续控制四轮驱动器无人机。我们的模型预测可解释的位置 - 访问分布，指示代理在执行过程中应在哪里进行以及应停止何处，并使用预测的分布来选择要执行的操作。这种两步模型的分解允许使用监督学习和模仿学习的组合进行简单有效的培训。我们通过逼真的无人机模拟器评估了我们的方法，并在两种最新的遵守方法中证明了绝对的任务完成精度提高了16.85％。
We propose an approach for mapping natural language instructions and raw observations to continuous control of a quadcopter drone. Our model predicts interpretable position-visitation distributions indicating where the agent should go during execution and where it should stop, and uses the predicted distributions to select the actions to execute. This two-step model decomposition allows for simple and efficient training using a combination of supervised learning and imitation learning. We evaluate our approach with a realistic drone simulator, and demonstrate absolute task-completion accuracy improvements of 16.85% over two state-of-the-art instruction-following methods.