作者: Angus Leigh , Joelle Pineau , Nicolas Olmedo , Hong Zhang
DOI: 10.1109/ICRA.2015.7139259
关键词: Artificial intelligence 、 Engineering 、 Robot operating system 、 Robot control 、 Mobile robot 、 Computer vision 、 Tracking (particle physics) 、 Human–computer interaction 、 Software 、 Robot 、 Laser data 、 Wheelchair
摘要: Having accurate knowledge of the positions people around a robot provides rich, objective and quantitative data that can be highly useful for wide range tasks, including autonomous person following. The primary this research is to promote development robust, repeatable transferable software robots automatically detect, track follow in their environment. work strongly motivated by need such functionality onboard an intelligent power wheelchair designed assist with mobility impairments. In paper we propose new algorithm robust detection, tracking following from laser data. We show approach effective various environments, both indoor outdoor, on different platforms (the Clearpath Husky). method has been implemented Robot Operating System (ROS) framework will publicly released as ROS package. also describe release several datasets standardized evaluation similar algorithms.