Takemitsu Mori

Online collision and occlusion detection based on 3D Cuboid Maps for a daily assistive robot with a tilting LRF

This paper describes about online collision and occlusion detection for a daily assistive robot. It is assumed that 3D map about around environment is constructed by collecting datasets gradually measured using a Laser RangeFinder mounted on a tilting platform, we develop an approach to detect collision and occlusion based on the 3D map while a robot tries to manipulate objects in daily environment. As the map representation, Time-series Composite Cuboid Map is proposed. It copes with temporal sequence explicitly, and map updating is efficiently performed. Several experiments on the assumption of daily assistance by a robot, we show the effectiveness of our approach.

Detection of localization failure using logistic regression

Monte Carlo localization (MCL) is a sample-based approach for representing probability density for the pose of a robot. MCL is widely used for mobile robots because of its robustness with respect to sensor noise. On the other hand, MCL can fail to estimate a pose of a robot if objects block measuring range of a laser sensor of the robot. Even though MCL fails to estimate the pose of the robot, MCL does not stop to estimate the pose because MCL does not have a self-diagnostic function. Therefore, detection of localization failure is essential for a mobile robot application. In the present paper, we propose a novel approach that detects the localization failure of MCL through logistic regression. The proposed approach has two advantages. First, it can detect whether position errors is larger than 0.15 m with a high accuracy. Second, the proposed method can detect localization failure by means of a statistically modeled equation. Moreover, as an example of an application using the probability of localization failure, we have proposed a hybrid localization scheme with MCL and laser odometry. The practical effectiveness of the proposed scheme is verified through experiments.

Cuboid-based workspace mapping and plane detection using time-series range data

This paper describes methods of environment map representation and plane detection for daily assistive robots. Under the assumption that 3D information of around environment is constructed by collecting range data gradually measured, we propose an effective map representation named TCCM (Time-series Composite Cuboid Map). The method copes with temporal sequence explicitly, and map updating is efficiently performed. In addition, two plane estimation methods are proposed. Through several experiments in daily environments, we ensured the effectiveness of our method.

Cuboid-based mapping using time-series range data for a daily assistive robot working on a daily environment

This paper describes a method of environment representation based on range data. The method named ‘Time-series Composite Cuboid Map’ is useful for daily assistive robots working on everyday environments. It explicitly copes with temporal sequence of range data, and contributes to online mapping about the following items: (1) Taking over the advantages of Multi-Level Surface Map, high power of expression with a small number of elements is achieved and (2) the compliance of temporal order of the environment map was kept according to addition and subtraction operation. The effective operation for map renewal is also proposed. As verification experiments, both a Laser RangeFinder mounted on tilting platform and a range camera are used, and mapping performance was confirmed. In addition, an application to collision and occlusion detection was implemented to a life-sized robot and the effectiveness of the proposed method was proved.