-
Tactile Aware Dynamic Obstacle Avoidance in Crowded Environment with Deep Reinforcement Learning
Authors:
Yung Chuen Ng,
Qi Wen,
Lim,
Chun Ye Tan,
Zhen Hao Gan,
Meng Yee,
Chuah
Abstract:
Mobile robots operating in crowded environments require the ability to navigate among humans and surrounding obstacles efficiently while adhering to safety standards and socially compliant mannerisms. This scale of the robot navigation problem may be classified as both a local path planning and trajectory optimization problem. This work presents an array of force sensors that act as a tactile laye…
▽ More
Mobile robots operating in crowded environments require the ability to navigate among humans and surrounding obstacles efficiently while adhering to safety standards and socially compliant mannerisms. This scale of the robot navigation problem may be classified as both a local path planning and trajectory optimization problem. This work presents an array of force sensors that act as a tactile layer to complement the use of a LiDAR for the purpose of inducing awareness of contact with any surrounding objects within immediate vicinity of a mobile robot undetected by LiDARs. By incorporating the tactile layer, the robot can take more risks in its movements and possibly go right up to an obstacle or wall, and gently squeeze past it. In addition, we built up a simulation platform via Pybullet which integrates Robot Operating System (ROS) and reinforcement learning (RL) together. A touch-aware neural network model was trained on it to create an RL-based local path planner for dynamic obstacle avoidance. Our proposed method was demonstrated successfully on an omni-directional mobile robot who was able to navigate in a crowded environment with high agility and versatility in movement, while not being overly sensitive to nearby obstacles-not-in-contact.
△ Less
Submitted 19 June, 2024;
originally announced June 2024.
-
Navigation with Tactile Sensor for Natural Human-Robot Interaction
Authors:
Zhen Hao Gan,
Yangwei You,
Meng Yee,
Chuah
Abstract:
Tactile sensors have been introduced to a wide range of robotic tasks such as robot manipulation to mimic the sense of human touch. However, there has only been a few works that integrate tactile sensing into robot navigation. This paper describes a navigation system which allows robots to operate in crowded human-dense environments and behave with socially acceptable reactions by utilizing semant…
▽ More
Tactile sensors have been introduced to a wide range of robotic tasks such as robot manipulation to mimic the sense of human touch. However, there has only been a few works that integrate tactile sensing into robot navigation. This paper describes a navigation system which allows robots to operate in crowded human-dense environments and behave with socially acceptable reactions by utilizing semantic and force information collected by embedded tactile sensors, RGB-D camera and LiDAR. Compliance control is implemented based on artificial potential fields considering not only laser scan but also force reading from tactile sensors which promises a fast and reliable response to any possible collision. In contrast to cameras, LiDAR and other non-contact sensors, tactile sensors can directly interact with humans and can be used to accept social cues akin to natural human behavior under the same situation. Furthermore, leveraging semantic segmentation from vision module, the robot is able to identify and, therefore assign varying social cost to different groups of humans enabling for socially conscious path planning. At the end of this paper, the proposed control strategy was validated successfully by testing several scenarios on an omni-directional robot in real world.
△ Less
Submitted 23 November, 2022;
originally announced November 2022.
-
Preparation and transport properties of non-hydrated Na$_{0.5}$CoO$_2$ single crystals
Authors:
X. Z. Chen,
Z. A. Xu,
G. H. Cao,
J. Q. Shen,
L. M. Qiu,
Z. H. Gan
Abstract:
Single crystals of Na$_{0.5}$CoO$_2$ were obtained through a flux method followed by de-intercalation of sodium. The Na$_{0.5}$CoO$_2$ samples were found to be vulnerable to water in the air and a hydration process in which H$_2$O molecules fill oxygen vacancies in CoO$_2$ layers is suggested to be responsible for the unusual vulnerability to water. The transport properties, including resistivit…
▽ More
Single crystals of Na$_{0.5}$CoO$_2$ were obtained through a flux method followed by de-intercalation of sodium. The Na$_{0.5}$CoO$_2$ samples were found to be vulnerable to water in the air and a hydration process in which H$_2$O molecules fill oxygen vacancies in CoO$_2$ layers is suggested to be responsible for the unusual vulnerability to water. The transport properties, including resistivity (\emph{$ρ$}), thermopower (\emph{S}) and Hall coefficient (\emph{R$_H$}), were studied in a temperature range of 5-300 K. The compound shows a weak localization of carriers just below 200 K and Co$^{3+}$-Co$^{4+}$ charge ordering at about 30 K, a relatively lower temperature than previously reported. The results seem to be quite different from those previously reported for this system [Foo et al, Phys. Rev. Lett. 92 (2004) 247001]. Possible mechanism underlying this kind of inconsistency is discussed.
△ Less
Submitted 11 December, 2004;
originally announced December 2004.
