[2009.07476] Path Planning using Neural A* Searchopen searchopen navigation menucontact arXivsubscribe to arXiv mailings

We present Neural A*, a novel data-driven search algorithm for path planning problems. Although data-driven planning has received much attention in recent years, little work has focused on how search-based methods can learn from demonstrations to plan better. In this work, we reformulate a canonical A* search algorithm to be differentiable and couple it with a convolutional encoder to form an end-to-end trainable neural network planner. Neural A* solves a path planning problem by (1) encoding a visual representation of the problem to estimate a movement cost map and (2) performing the A* search on the cost map to output a solution path. By minimizing the difference between the search results and ground-truth paths in demonstrations, the encoder learns to capture a variety of visual planning cues in input images, such as shapes of dead-end obstacles, bypasses, and shortcuts, which makes estimated cost maps informative. Our extensive experiments confirmed that Neural A* (a) outperformed

1 mentions: @RYonetani
Date: 2020/09/17 02:21

Referring Tweets

@RYonetani arXivに新しい論文をアップロードしました: "Path Planning using Neural A* Search" t.co/4e0Sxvtfrn 米谷・谷合・西村(オムロンサイニックエックス)、Amin(インターン)、技術アドバイザの金崎先生(東工大)による研究で、微分可能なA*探索による経路計画の話です。

Bookmark Comments

Related Entries

Read more [1811.02091] Simple, Distributed, and Accelerated Probabilistic Programming
4 users, 0 mentions 2018/12/02 15:01
Read more GitHub - champierre/ml2scratch: Connect Machine Learning(TensorFlow.js) with Scratch.
1 users, 1 mentions 2019/08/11 14:16
Read more [2001.07895] Partially-Shared Variational Auto-encoders for Unsupervised Domain Adaptation with Targ...
0 users, 3 mentions 2020/07/02 23:21
Read more 驥大エ� 譛晏ュ�
0 users, 1 mentions 2020/07/03 02:21
Read more Paper accepted at ECCV2020 - sinicx - Medium
0 users, 2 mentions 2020/07/08 01:16