The present paper describes the integration of laser-based perception, footstep planning, and walking control of a humanoid robot for navigation over previously unknown rough terrain. A perception system that obtains the shape of the surrounding environment to an accuracy of a few centimeters is realized based on input obtained using a scanning laser range sensor. A footstep planner decides the sequence of stepping positions using the obtained terrain shape. A walking controller that can cope with a few centimeters of error in terrain shape measurement is achieved by combining the generation of a 40-ms cycle online walking pattern and a ground reaction force controller with sensor feedback. An operational interface was developed to send commands to the robot. A mixed-reality display was adopted to realize an intuitive interface. The navigation system was implemented on the HRP-2, a full-size humanoid robot. The performance of the proposed system for navigation over unknown rough terrain and the accuracy of the terrain shape measurement were investigated through several experiments.

This paper describes a method for tracking of moving objects in multiple categories with a LRF (laser range finder). The proposed method can track multiple moving objects robustly using a framework based on SJPDAF (sample-based joint probabilistic data association filter). It can continue tracking correctly if moving objects get close to other moving objects, and can track moving objects in different categories at the same time. We implemented the proposed method and obtained experimental results demonstrating its effectiveness.

Autonomous mobile robots are required to understand surrounding environments for its localization and motion planning. This paper proposes a 3D mapping system using LIDAR and a triaxial inertial sensor. The system is based on "Velodyne SLAM" which is a state of the art ICP based mapping method using only point cloud data. In comparison to the existing method, the proposed method is robust to rotary motion and works for fast and large change of sensor position and orientation. We implemented the proposed method to the hand-held sensor unit, and evaluated its effectivity in varied environments.

This paper presents an online incremental method for object discovery by a mobile robot equipped with a LiDAR sensor. Obtaining object information from the real world is important for an autonomous mobile robot to enable it to understand its environment. The paper focuses on gathering object information as a point cloud from a mobile robot. In contrast to existing systems using dense 3D data, we propose an online incremental system using sparse point cloud data at each frame. Experiment results show the system effectively extracts surrounding objects.

The paper presents a path planning method for an autonomous mobile robot working in human environment. The traditional planning method uses the map having only static obstacle information. A path generated by that method does not consider environmental rules. For example no trespassing area or one side-way passage. To solve this problem, we propose a path planning method with pedestrian information map. We define pedestrian as human moving to the destination. We generate pedestrian information map using only pedestrian traffic information in all observed human traffic information. Pedestrian information map shows passable areas and the traffic manner on the map. Experimental results showed the effectiveness of the proposed method.

Analyzing the auditory scene of real environments is challenging partly because an unknown number and type of sound sources are observed at the same time and partly because these sounds are observed on a significantly different sound pressure level at the microphone. These are difficult problems even with state-of-the-art sound source localization and separation methods. In this paper, we exploit two such methods using a microphone array: (1) Bayesian nonparametric microphone array processing (BNP-MAP), which is capable of separating and localizing sound sources when the number of sound sources is unspecified, and (2) robot audition software “HARK” is capable of separating and localizing in real time. Through experimentation, we found that BNP-MAP is more robust against differences in the sound pressure levels of the source signals and in the spatial closeness of source positions. Experiments analyzing real scenes of human conversations recorded in a big exhibition hall and bird calling recorded at a natural park demonstrate the efficacy and applicability of BNP-MAP.

This paper describes a method for classifying and tracking multiple moving objects with a laser range finder (LRF). As moving objects are tracked in the framework of sample-based joint probabilistic data association filters (SJPDAFs), the proposed method is robust against occlusions or false segmentation of LRF scans. It divides tracking targets and corresponding LRF segments into clusters and able to classify each cluster as a car or a group of pedestrians. In addition, it can correct false segmentation of LRF scans. We implemented the proposed method and obtained experimental results demonstrating its effectiveness in outdoor environments and crowded indoor environments.

This paper describes environmentmaps that are comprised of the following three types of information, 1) 3D environmental changes that represents human activities, 2) human trajectories in 2D that represent how humans move in the environment, and 3) human posture data. These maps are utilized in order to plan safer, quicker and/or non-human-disturbingpaths for a mobile service robot at themuseum “Miraikan”. Experiments are conducted within “Miraikan” and results are shown.

Tracking method of moving objects using laser range finders can be applied to various applications. However, it is inefficient to make separate programs for each application, and it also causes bugs. This paper describes the designing of our versatile moving object tracking platform applied to several applications.

The paper presents an object identification method that can identify known and unknown object with its position from 3D point cloud data. The proposed system is divided into two main steps: LIDAR based segmentation and extracted objects identification. LIDAR sensor gives sparse but accurate 3D shape information compared to stereo camera, and it covers omnidirectional larger area. One of main problem of using such data is that the object shape information is varied by its orientation and distance from the sensor. The conventional identification model needs to specify a fixed number of classes and applicable dimensions to describe an object in advance. To solve the problem, we use the nested infinite Gaussian mixture model for object identification. The experimental results show that the proposed system could extract varied objects and identify known and unknown objects including segmentation errors.

We propose a method of extracting watermark from 3D prints created from 3D mesh data. The watermark is embedded to the 3D mesh in the spectral domain using a robust, imperceptible, and informed algorithm based on the spread spectrum technique, and then extracted from 3D prints by reconstructing the 3D mesh homologous to the original. A suspect 3D mesh is reconstructed accurately and robustly from noisy and incomplete 3D scans of the 3D prints, by optimizing a sparse subset of the spectrum using a variant of the iterative closest point technique. The mesh is registered to the original mesh by construction, in terms of geometry and topology, which allows us to extract a suspect watermark by simple algebraic operations. Comprehensive experiments have been conducted to demonstrate the performance of our method using standard 3D datasets and multiple 3D printers. The proposed method significantly outperforms prior methods under different conditions in the practical scenario. The probability of false positive detection is kept less than 106 for the most cases of simulated and real data.

Localisation and mapping are fundamental capabilities for autonomous mobile robots, and there has been a large amount of recent work in these fields. However, much of the work does not consider dynamic environments that include humans and moving objects. Such objects can cause occlusions resulting in a fewer visible landmarks, which can decrease localisation performance. This paper describes a novel method of localisation and multi-layered 2D mapping in dynamic environments using selective updating of a particle filter. A number of horizontal, planar laser scans at varying heights are used to construct a number of corresponding 2D maps. At each mapping step, the position estimate from the map layer which minimizes uncertainty is selected and used to update all maps. Using the proposed method, it is possible to localize precisely in dynamic environments, despite the effects of occlusion. Experimental results in a large outdoor environment confirms the effectiveness of the method.

Legged robots have unique capabilities to traverse complex environments by stepping over and onto objects. Many footstep planners have been developed to take advantage of these capabilities. However, legged robots also have inherent constraints such as a maximum step height and distance. These constraints typically limit their reachable space, independent of footstep planning. Thus, we propose that robots such as humanoid robots that have manipulation capabilities should use them. A robot should autonomously modify its environment if necessary. We present a system that enabled a real robot to use a box to create itself a stair step or place a board on the ground to cross a gap, allowing it to reach its otherwise unreachable goal configuration.

For an autonomous mobile robot that works in real-world environment, recognition of its surrounding environment is necessary. This paper presents an object identification system, which can identify known and unknown objects and estimate their locations using 3D point cloud data acquired from a 3D LIDAR sensor mounted on the mobile robot. The proposed system is divided into two main steps; the first step is segmentation based on the 3D point cloud and the second step is identification of the extracted objects. The 3D LIDAR sensor gives sparse 3D shape information accurately, and covers a wide range. One of the main problems of such data is that the object shape information varies according to object's orientation and distance from the sensor. To solve this problem, we use nested infinite Gaussian mixture models for object identification. The experimental results show that the proposed system can extract various types of objects and identify both known and unknown objects.

This paper demonstrates some experimental proofs of the model for the classification and identification of robot sensing data. Autonomous robots are equipped with varied sensors to assist them in understanding and interacting with their environments. In contrast to traditional model approaches that are based on the Gaussian assumption, we propose the application of the infinite Gaussian mixture model (iGMM) to detect known and unknown data. Two key components are denoted: 1) simultaneous training of the number of classes and dimensions of each model, and 2) infinite modeling to adjust for observations that do not match with previous knowledge.

As a first step, an autonomous robot needs to know where is himself in the environment. Simultaneous Localization and Mapping (SLAM) is a method of estimating its own position and creating a map of the surrounding environment in unknown area at the same time. In the conventional SLAM method, a wheel rotation angle (wheel odometry) has been need, but it cause some problems, the method is vulnerable to tilt and slip, difficult to porting to other robots, especially, the robot without wheels. Therefore, we propose the method of SLAM using IMU-based Visual Odometry and 3D-LIDAR. We use 6DoF estimation from a visual odometry instead of wheel odometry input, and to generate one-line laser scan from 3D-LIDAR. As a result, we obtain the real-time SLAM without wheel odometry, and demonstrate that it can apply not only to wheel mobile robot but also to human hand-held.

Recently robotic localization and mapping have been a field of study. They are the basic function of a mobile robot. However such methods are not enough in the dynamic environment including human and moving objects. In such field, there are serious problems such as occlusion and fewer landmarks, that make localization error. This paper describes a nobel method of localization and multi-layered 2D mapping in the dynamic environment using selective update for particle filter. It is possible to localize precisely, in spite of negative factor in the dynamic environment. Also we can exactly make multi-layered 2D maps.