This paper introduces a context-aware assisted active robotic walker for Parkinson's disease (PD) patients. Most of PD patients suffer from not only loss balance but also abnormal gaits. These symptoms tend to make fall down more easily and result in low quality life. We use Hidden Markov Model (HMM) analyze the gait patients, then our help adjust their become normal while applying auditory cues when gaits are recognized. To prevent user leaning forward before falling down, locks motors...

In this paper, we propose a sensory cues guided robotic walker for improving gaits of Parkinson Disease (PD) patients. A completely non-intrusive, 3D real-time leg pose tracking and gait analysis are proposed by using depth camera mounted on the rear walker. It has been studied that can serve as effective stimuli to PD patients improvement. our work, including visual auditory incorporated into More specifically, both gait-adaptive, which cue in particular is projected onto ground projector...

This paper presents a hybrid system of applying homeostatic drives theory and stimuli from environment or human on robot, aiming to solve the increasing need elder caring in modern aging society. Through reinforcement learning with feedback, robot is enabled make appropriate decisions by reviewing its own motivation, inputs, allowing finish vague tasks, i.e. serve, help have conversation different situations. In our experiment, satisfaction rate could reach 94% as well maintain stable...

For a service robot, it is not adequate to let its navigational movement be based only on single metric, such as minimum distance path. In the environment where robot and humans are coexisting, should always perform social navigation whenever moving. However, navigation, needs follow certain "social norms" of environment. Recently, deep reinforcement learning (DRL) technique popularly applied robotics field; yet, rarely used solve mentioned problem, generally deemed high dimension complex...

People engaging in an activity usually has individual tolerance to be interrupted [1], [2]. Humans subconsciously adapt their behaviors draw other one's attention and get into a conversation based on historical experiences, but robots often fail aware of humans' feeling thus interrupt users repeatedly. To endow service with such socially acceptable ability, we propose online human-aware interactive learning framework this paper, under which the robot personalizes its according both observed...

In this paper, we present a novel data structure representing the environment with occupancy grid cells while each map is associated set of line features extracted from laser scan points. Due to fact that segments are principal elements artificial environments, they provide considerable geometric information about which can be used for enhancing accuracy localization. Orthogonal characteristic key issue guarantee consistency SLAM algorithm by allowing us deal lines parallel or perpendicular...

As robots are put into humans' daily life, the assigned tasks to varied, and different needs of people interacting with immense. a result, when facing users, it is important for personalize interactions provide user-desired services. This paper, therefore, proposes learning strategy on service-providing model. Through human feedback, enables robot learn users' needs, as well preferences, adjust its behaviors. Here, we assume that preferences may vary time; hence goal this paper let...

This paper presents our development of a new outdoor tour guide robot, named, NTU-I. It is constructed to provide autonomous guiding services on the campus National Taiwan University. In order fulfill reliability and safety navigation, data acquired via several sensing technologies, such as differential global positioning system (DGPS), dead reckoning, digital compass are fused by way well known extended Kalman filtering (EKF) technique. Furthermore, shortest path planning an obstacle...

For service robots to be able enter a multi-human office environment, it is important find group of human users' social patterns and then provide proper them in time. Usually, are represented terms nonverbal signals. In this paper, new integrated approach on recognizing signals proposed. Specifically, the detected by laser range finder RGB-D camera processed (spatial) patterns. Those recognized applied human-to-human, human-to-robot or multi-human-to-robot interactive formation. Experimental...

RGB-D sensor has gained its popularity in the study of object recognition for low cost as well capability to provide synchronized RGB and depth images. Thus, researchers have proposed new methods extract features from data. On other hand, learning-based feature representation is a promising approach 2D image classification. By exploiting sparsity signals, we can learn instead using hand-crafted local descriptors like SIFT or HoG. This framework inspired us Our work focuses on two goals....

In an ageing society, we expect that a robotic caregiver is able to persuade the elderly perform healthier behavior. this work, pragmatic argument adopted make realize choice beneficial for health really worthwhile, such as eating suitable fruits. Based on concept, adaptive recommendation dialogue system through argumentation proposed. There are three objectives in system. First, knowledge base construction built, which concerns not only effect of decision but also reason effect. Secondly,...

In this paper, we propose an online scene reconstruction algorithm with monocular camera since there are many advantages on modeling and visualization of environment physical instead resorting to sparse 3D points. The goal is simultaneously track the position map environment, which close spirit visual SLAM. There're plenty SLAM algorithms in current literature can provide a high accuracy performance, but them rely stereo cameras. It's true that we'll face more challenges accomplish task...

For a visual servo system, there usually exists the problem of time-delay likely caused by long image processing and data transmission. The system our robot is subject to two main limitations stemming from specific commercial mobile manipulator, which one large due transmission, whereas other failure directly command each joint velocity as servo's control input. In this paper, we propose novel compensate time-delay, use integral incorporating estimation prediction with kinematic model so...

In this paper, we propose a framework to perceive the level of intimacy in dyadic human interactions from both robot perspective and first-person perspective. First all, for catching insight interaction with three different degrees persons, namely normal, familiar close, have done preliminary user study social interaction. Next, field science our study, design four types features, consisting proxemics, non-verbal, verbal, temporal features categorize observations, but including them reaches...

To enable service robots enter a multi-human office environment, it is important to find group of human users' social patterns and then provide proper them in time. Usually, are represented terms nonverbal signals. In this paper, some signals fast detected environments. Then, robot can the spatial users. Those indicated human-to-human, human-to-robot or multi-human-to-robot interaction. Experimental results shows that our successfully aforementioned patterns.

In a situation where robot initiates interaction with groups of people, questions such as "where is the people" should be addressed. This paper proposed real-time method that enables to find position and orientation people. The mainly fuse depth-related data track positions people compute orientations those by using data. conducted experimental results demonstrate properness in knowing are.