Kok Seng Eu

3D Smart User Interactive System with Real-Time Responding Tele-Robotic Proprioceptive Information

Feedback of proprioceptive information is essential for many tele-robotic systems, especially those designed to undertake tasks concerning hazardous environments and for efficient out-of-sight remote control applications. Given highly sensitive nature of these applications, even small errors (e.g. less than one degree of displacement in robot posture) can cause unnecessary risk. Thus, accurate feedback of proprioceptive information, as well as a technique to precisely interpret this information, is significant to operator. In this paper, we introduce a framework that uses pulse feedback mechanism to measure the proprioceptive information of a robot operating over real-time wireless communication and represent it in 3D model user interface. The 3D user interface enhances the interpretation of proprioceptive information to help operator to visualize the real-time relative position of the robot. The paper also provides results that demonstrate how the framework allows synchronization between 3D model and tele-robot to be achieved in real-time over wireless communications.

An Airflow Analysis Study of Quadrotor Based Flying Sniffer Robot

Wheeled based sniffer robots have been using in current research trend of odor plume tracking, but they have a serious pitfall because they only perform2D odor plume tracking which is ineffective. The main reason of being ineffective is because wheeled sniffer robots ignore the fact that the majority of odor plumes are released into 3D space instead of 2D space. Therefore, a flying sniffer robot is needed to overcome 3D odor plume tracking problem. In this paper, we propose to use a quadrotor as the locomotion of a flying sniffer robot. In order to make sure the quadrotor based flying sniffer robot can perform well in odor plumes tracking, we have carried out a detail study in airflow simulation analysis by using Computational Fluid Dynamics (CFD) software. Besides, we also conducted experiments to determine the odor sensing ability of quadrotor based flying sniffer robot. From our experiments, we are able to identify the odor sensing region of quadrotor based flying sniffer robot in quantitative measurement. From airflow analysis simulation study and experiments, we proved that a quadrotor based flying sniffer robot is a feasible solution for 3D odor plume tracking.

A network-adaptive compensation technique for tele-haptics using position prediction algorithm

Haptic interfaces have been applied as controllers in many areas especially tele-operation and distributed virtual environments. They are used to manipulate objects in both physical and virtual environments. Haptics enhance force interactions and provides a better immersive user experience. Moreover haptic devices inherently function in close proximity to humans. In the case of network-based haptic control, the network criteria for stability in haptic interactions are much more sophisticated than that for tethered devices and the kinematic and force data are required to be transferred over communication link within a stringent time. Otherwise, users may feel unexpected vibration or abrupt force even when the device is in free motion. This is mainly due to position de-synchronization between local and remote environments. This paper addresses position synchronization which is the essential problem to lose stability of haptic experience under the influence of network and system impairments. A novel encoder referencing position synchronization algorithm is proposed in order to compensate the network latency and packet loss over network. This has been successfully tested under influence of the network impairment and low transmission rates.

Chemical plume tracing: A three-dimensional technique for quadrotors by considering the altitude control of the robot in the casting stage

Most of the reported three-dimensional chemical plume tracing methods use stereo sensing method to determine the next tracing step direction. For example, multiple sensors are used for detection in the left, right, up and down directions. Left and right detections are feasible for stereo sniffing; unfortunately, the same approach is infeasible for the up/down sensing of the quadrotor platforms because the propellers of the quadrotor continuously draw the air from the top and bring it down, which affects the sensing of the upper and lower sensors, and fails to determine the subsequent tracing step of up/down direction. Therefore, up/down sensing in the surging stage of chemical plume tracing is ineffective for quadrotor platforms (chemical plume tracing has two stages: surging and casting). To solve the problem, we propose an alternative that is not in the surging but in the casting stage of chemical plume tracing, by designing a new three-dimensional chemical plume tracing technique with variations of altitude (z-axis) control during the casting stage, which has never been considered in the previous works. Besides, we use a computational fluid dynamics software to study the airflow pattern of quadrotor platform. Subsequently, a fuzzy-based stereo-sniffing algorithm is developed by considering the quadrotor propeller’s air intake stream angle associated with the environmental wind direction angle, so as to improve the accuracy of stereo sensing. The results of the proposed solutions are verified and validated via both experimental and simulation approaches.

A quantitative evaluation of haptic data prediction techniques over best-effort network

Exchanging haptic information over best-effort networks such as the Internet presents challenges due to the extremely high sensitivity to network impairments, especially the simultaneous occurrence of time-varying network latency and packet loss. Subsequently, the haptic interaction experience is deteriorated along with a reduction in the fidelity. In this paper we present a new approach to mitigate the effects of network impairments, termed Trust Strategy Prediction. As well as evaluation of TSP in quantitative model is presented in terms of accuracy and smoothness of haptic updates and compared with two well-known techniques used in haptic data prediction: Dead Reckoning and velocity/yank Estimation with filtering mechanism.

A Simulation Study of Micro-Drone Chemical Plume Tracking Performance in Tree Farm Environments

Chemical plume tracking (CPT) technology is the mean of tracking the flow of specific chemical plume in the air, to locate the source. Nowadays, CPT technology, for instance, a micro-drone based chemical plume tracking robot, has great potential in identifying hidden explosives, illegal drugs and blood for police and military purposes. However, environmental factors such as obstacles on site can change the wind vectors will cause inconsistent odor plume propagation. With most of the previous work conducted from numerous researchers carried out in empty open space, this paper studies the influence of obstacles on site towards CPT’s performance, which the simulation focus in one specific environment, a tree farm, with different density of trees or trees’ spacing. For this paper, we developed a 3D gas dispersion simulator with mobile robot olfaction (MRO) capability. Through the simulation, correlation between the impacts of tree farm density factor to CPT’s performance is found out, where higher tree density (or smaller tree spacing distance) can significantly reduce the performance of CPT. This study is an important fundamental contribution for drone’s CPT operation in agriculture application beneficial to future use, such as smell tracking of mature fruits in tree farm.

Wind direction and speed estimation for quadrotor based gas tracking robot

In gas extraction sites, the incidents of gas leaking poses a damage to workers on site. The protection for the workers is essential. However, due to the colorless and odorless nature of natural gas, it is difficult for humans to identify leaks. This paper proposes a quadrotor based gas tracking robot to be used in hazardous gas localization areas, and specifically, to detect methane leak from gas extraction sites. For the quadrotor to fulfill this purpose, it requires the ability to detect wind direction and speed, an endowment that commercial quadrotors lack. The need to detect wind direction and speed stems from the fact that gas plumes travel downwind, but the quadrotor needs to find the source of the leak, and hence, must determine the upwind direction to locate the source. In order to equip the quadrotor with the above skills, a wind direction and speed estimation algorithm based on Euler angles-velocity vectors has been proposed. For comparison purposes, we compared the proposed method with a generic ultrasonic wind sensor. We concluded that the proposed method achieves an error percentage as low as 10.73% for wind speed, and 9.09% for wind direction estimation. Thus, the algorithm is a significant addition to the quadrotors’ capabilities, enabling the quadrotor to trace upwards, against the traversal of the gas plume, and carry out accurate calculations.

Single Beacon Indoor Localization System Based on Counter-Synchronized Compass and RSSI

Indoor localization system has been a popular research area in recent decades and many of them are based on multiple beacons localization method. However, there are some special applications to which the multiple beacons method is not an optimal solution due to its overdesign and cost of redundancy. Multiple beacons method uses at least three transducers and each transducer’s location must be known to find the location of a target object by using either Triangulation or Trilateration calculation. When the multiple beacons method is applied in an items lost and found system, the precise Cartesian coordinates of a target item can be found, but it is definitely overdesign and incurring redundant cost. It is due to the fact that the target item requires only two simple information i.e. Clock orientation and distance information; therefore, single beacon is enough for the task. In this paper, we propose a single beacon localization method to optimize the solution in the items lost and found system by utilizing clock orientation and estimated distance information. The proposed single beacon localization algorithm has been demonstrated and proven that it can be one of the optimal solutions for items lost and found system.

3D smart user interactive system with proprioceptive orientation feedback by using Disparity Compassing

Feedback of proprioceptive information is important for many tele-robotic systems, especially performing a navigation task and out-of-sight remote controlling. Proprioception is defined as the ability to sense the position, location, orientation, and movement of the body and its parts. Although proprioceptive feedback is one of the significant feedbacks for tele-operations, it has not been taken much consideration in current tele-robotics research. This paper emphasizes on the orientation feedback of a robot with a 3D model user interface (UI). The 3D UI enhances the interpretation of proprioceptive information in order to aid operator to visualize the relative position and orientation of the robot. A new method called “Disparity Compassing” is proposed and applied into tele-robotic framework by using two magnetic compass sensors. One of the sensors is mounted on a robot and the other one is built-in compass sensor of the UI. With the disparity values of two compass sensors, the interactive orientation information of the robot can be fed back in real time to the operator. In this paper, the disparity compassing technique has been demonstrated. It has been proven that the tele-robotic system with disparity compassing enhances the awareness of relative positional navigation in the situation of unknown space exploration.