Supermedia-enhanced Internet-based telerobotics

This paper introduces new planning and control methods for supermedia-enhanced real-time telerobotic operations via the Internet. Supermedia is the collection of video, audio, haptic information, temperature, and other sensory feedback. However, when the communication medium used, such as the Internet, introduces random communication time delay, several challenges and difficulties arise. Most importantly, random communication delay causes instability, loss of transparency, and desynchronization in real-time closed-loop telerobotic systems. Due to the complexity and diversity of such systems, the first challenge is to develop a general and efficient modeling and analysis tool. This paper proposes the use of Petri net modeling to capture the concurrency and complexity of Internet-based teleoperation. Combined with the event-based planning and control method, it also provides an efficient analysis and design tool to study the stability, transparency, and synchronization of such systems. In addition, the concepts of event transparency and event synchronization are introduced and analyzed. This modeling and control method has been applied to the design of several supermedia-enhanced Internet-based telerobotic systems, including the bilateral control of mobile robots and mobile manipulators. These systems have been experimentally implemented in three sites test bed consisting of robotic laboratories in the USA, Hong Kong, and Japan. The experimental results have verified the theoretical development and further demonstrated the stability, event transparency, and event synchronization of the systems.     

Design and Analysis of Internet-Based Tele-Coordinated Multi-Robot Systems

The coordination of multi-robots is required in many scenarios for efficiency and task completion. Combined with teleoperation capabilities, coordinating robots provide a powerful tool. Add to this the Internet and now it is possible for multi-experts at multi-remote sites to control multi-robots in a coordinated fashion. For this to be feasible there are several hurdles to be crossed including Internet type delays, uncertainties in the environment and uncertainties in the object manipulated. In addition, there is a need to measure and control the quality of tele-coordination. To this end, the measure of force sensed by each robot is suggested and justified as a coordination index. It was proven that if n robots are event-transparent and event-synchronous then they can be teleoperated under random delay conditions to coordinate to any index value, which is feasible under no delay conditions. The design procedure that ensures a system can satisfy a small coordination index was presented and analyzed.In addition, the design and analysis of event-synchronous systems using Petri Nets is detailed. The Petri Net design methodology is presented for both event-synchronous single operator single robot teleoperation systems and event-synchronous multi-operator multi-robot teleoperation systems.The theory developed was tested by bilaterally tele-coordinating two mobile manipulators via the Internet. The experimental results confirmed the theoretical results presented.     

Dolphins (Tursiops truncatus) comprehend the referential character of the human pointing gesture.

The authors tested a dolphin's (Tursiops truncatus) understanding of human manual pointing gestures to 3 distal objects located to the left of, to the right of, or behind the dolphin. The human referred to an object through a direct point (Pd), a cross-body point (Px), or a familiar symbolic gesture (S). In Experiment 1, the dolphin responded correctly to 80% of Pds toward laterally placed objects but to only 40% of Pds to the object behind. Responding to objects behind improved to 88% in Experiment 2 after exaggerated pointing was briefly instituted. Spontaneous comprehension of Pxs also was demonstrated. In Experiment 3, the human produced a sequence of 2 Pds, 2 Pxs, 2 Ss, or all 2-way combinations of these 3 to direct the dolphin to take the object referenced second to the object referenced first. Accuracy ranged from 68% to 77% correct (chance?=?17%). These results established that the dolphin understood the referential character of the human manual pointing gesture. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Haptic information in Internet-based teleoperation

Many tasks can be done easily by humans turn out to be very difficult to accomplish with a teleoperated robot. The main reason for this is the lack of tactile sensing, which cannot be replaced by visual feedback alone. Once haptic devices are developed, their potential in many fields is obvious. Especially, in teleoperation systems, where haptic feedback can increase the efficiency and even render some tasks feasible. This paper studies Internet-based teleoperation systems that include haptic feedback, concentrating on the control of such systems and their performance. The potential of this technology and its advantages are explored. In addition, key issues, such as stability, synchronization, and transparency are analyzed and studied. Specifically, an event-based planning and control of Internet-based teleoperation systems is presented with experimental results of several implemented system scenarios in micro- and macro-scales     

Data representation for CNN based internet traffic classification: a comparative study

Multimedia Tools and Applications - It has been well established that the Internet of Things will bring an expansion in traffic volume and types. This will bring new challenges in terms of Quality...     

Personalized teleoperation via intention recognition

ABSTRACT

One of the challenges of teleoperation is the recognition of a user’s intended commands, particularly in the manning of highly dynamic systems such as drones. In this paper, we present a solution to this problem by developing a generalized scheme relying on a Convolutional Neural Network (CNN) that is trained to recognize a user’s intended commands, directed through a haptic device. Our proposed method allows the interface to be personalized for each user, by pre-training the CNN differently according to the input data that is specific to the intended end user. Experiments were conducted using two haptic devices and classification results demonstrate that the proposed system outperforms geometric-based approaches by nearly 12%. Furthermore, our system also lends itself to other human–machine interfaces where intention recognition is required.     

Decision confidence-based multi-level support vector machines

Support vector machines (SVM) have been showing high accuracy of prediction in many applications. However, as any statistical learning algorithm, SVM's accuracy drops if some of the training points are contaminated by an unknown source of noise. The choice of clean training points is critical to avoid the overfitting problem which occurs generally when the model is excessively complex, which is reflected by a high accuracy over the training set and a low accuracy over the testing set (unseen points). In this paper we present a new multi-level SVM architecture that splits the training set into points that are labeled as ‘easily classifiable’ which do not cause an increase in the model complexity and ‘non-easily classifiable’ which are responsible for increasing the complexity. This method is used to create an SVM architecture that yields on average a higher accuracy than a traditional soft margin SVM trained with the same training set. The architecture is tested on the well known US postal handwritten digit recognition problem, the Wisconsin breast cancer dataset and on the agitation detection dataset. The results show an increase in the overall accuracy for the three datasets. Throughout this paper the word confidence is used to denote the confidence over the decision as commonly used in the literature.     

EVoIP: Energy efficient voice over IP privacy

Due to the convergence of telecommunication technologies and pervasive computing, voice is increasingly being transmitted over IP networks, in what is commonly known as Voice over IP (VoIP). Despite many advantages offered by this technology, VoIP applications inherit many challenging characteristics from the underlying IP network related to quality of service and security concerns. Traditional ways to secure data over IP networks have negative effects on real-time applications and on power consumption, which is scarce in power-constrained handheld devices. In this work, a new codec-independent Energy Efficient Voice over IP Privacy (E²VoIP²) algorithm is devised to limit the overhead of the encryption process, without compromising the end-to-end confidentiality of the conversation. The design takes advantage of VoIP stream characteristics to encrypt selected packets using a secure algorithm, while relaxing the encryption procedure in-between these packets. We evaluated experimentally the difficulty of conducting known plaintext attacks on VoIP by demonstrating that a sound recorded simultaneously by different sources results in apparently random encoded files. Regarding E²VoIP², experimental and simulation results show a substantial improvement in terms of the number of CPU cycles which results in a reduction of latency and a reduction in consumed power with respect to that of the SRTP. In addition, the proposed method is flexible in terms of the balance between security and power consumption.     

Mobile Apps identification based on network flows

Network operators and mobile carriers are facing serious security challenges caused by an increasing number of services provided by smartphone Apps. For example, Android OS has more than 1 million Apps in stores. Hence, network administrators tend to adopt strict policies to secure their infrastructure. The aim of this study is to propose an efficient framework that has a classification component based on traffic analysis of Android Apps. The framework differs from other proposed studies by focusing on identifying Apps traffic from a network perspective without introducing any overhead on subscribers smartphones. Additionally, it involves a technique for pre-processing network flows generated by Apps to acquire a set of features that are used to build an identification model using machine learning algorithms. The classification model is built using classification ensembles. A group of chosen users contribute in training the classification model, which learns the normal behavior of selected Apps. Eventually, the model should be able to detect abnormal behavior of similar Apps across the network. A 93.78% classification accuracy is achieved with a low false positive rate under 0.5%. In addition, the framework is able to detect abnormal flows of unknown classes by implementing an outlier detection mechanism and reported a 94% accuracy.