Path planning is an essential function in autonomous vehicles. Several Reinforcement Learning (RL) algorithms were explored for achieving the path planning task. Challenges include generating optimal vehicle paths while maintaining high safety standards and low training time. In this paper, we present a Twin Delayed Deep Deterministic Policy Gradient (TD3) RL algorithm as a modular path planner. This path planner utilizes an occupancy grid map training environment, integrated with the Intel Responsibility Sensitive Safety (RSS) rules for planning high safety standards paths. The resulting TD3 algorithm is compared to its predecessor, the Deep Deterministic Policy Gradient (DDPG) RL algorithm, in terms of the training time. The results show that our proposed path planner successfully reduced the needed training time by 31% in comparison to the DDPG-based path planner. Additionally, CARLA simulations were provided for validation, showing the high safety standards in the planned paths.

Drivable area segmentation is an important feature for autonomous driving. Currently, state of the art techniques in this area focus on segmenting roads in urban areas with near perfect conditions. Roads with deteriorating conditions have received much less attention, even though they are common and present a unique set of challenges to the road segmentation tasks. These challenges include detecting obstacles (manholes and potholes) and determining whether or not it is safe to drive over them, and detecting road boundaries while lacking proper markings. This paper proposes a new method for drivable area segmentation in roads with deteriorating conditions based on a 3D LiDAR. Our framework represents the LiDAR point cloud data in an angular grid object which splits the data into smaller point cloud objects based on the laser scan number and the projection angle of each point. We apply multiple filtration steps in our framework in order to accurately detect the road boundaries and to detect and classify any road irregularities. The experiments on our collected datasets demonstrate the performance of our framework in detecting and classifying road drivable regions accurately and robustly. We reached a maximum precision of 92.78% in detection road boundaries and a max precision of 99.38% in detection and classifying road irregularities.

The cybersecurity community has always sought low-latency high-throughput anonymous systems. A common problem between all approaches to the problem of anonymity is the discovery process of an anonymous path between the sender and receiver. Currently, the most robust model for anonymous path discovery is TOR directory servers. Despite being distributed, TOR directory servers force clients to expose themselves to a potentially malicious directory. In this paper, we present KDST (K-Anonymous Node Discovery Using Separation of Trust), a model that provides a methodology to have true anonymous path discovery. The model is based on the separation of trust, and thus exposing the client would require full compromise of the system. Even with a fully compromised system, our approach incorporates K-anonymity to frustrate such de-anonymization attacks. We implement our model and assess its performance and security. Our analysis shows that KDST is very effective in solving the problem of anonymous node discovery.

Recent work showed that using image processing techniques on thermal images taken by high-end equipment reveals passwords entered on touchscreens and keyboards. In this paper, we investigate the susceptibility of common touch inputs to thermal attacks when non-expert attackers visually inspect thermal images. Using an off-the-shelf thermal camera, we collected thermal images of a smartphone's touchscreen and a laptop's touchpad after 25 participants had entered passwords using touch gestures and touch taps. We show that visual inspection of thermal images by 18 participants reveals the majority of passwords. Touch gestures are more vulnerable to thermal attacks (60.65% successful attacks) than touch taps (23.61%), and attacks against touchscreens are more accurate than on touchpads (87.04% vs 56.02%). We discuss how the affordability of thermal attacks and the nature of touch interactions make the threat ubiquitous, and the implications this has on security.

Eye-gaze and mid-air gestures are promising for resisting various types of side-channel attacks during authentication. However, to date, a comparison of the different authentication modalities is missing. We investigate multiple authentication mechanisms that leverage gestures, eye gaze, and a multimodal combination of them and study their resilience to shoulder surfing. To this end, we report on our implementation of three schemes and results from usability and security evaluations where we also experimented with fixed and randomized layouts. We found that the gaze-based approach outperforms the other schemes in terms of input time, error rate, perceived workload, and resistance to observation attacks, and that randomizing the layout does not improve observation resistance enough to warrant the reduced usability. Our work further underlines the significance of replicating previous eye tracking studies using today's sensors as we show significant improvement over similar previously introduced gaze-based authentication systems.

In this paper, we propose a calibration-free gaze-based text entry system that uses smooth pursuit eye movements. We report on our implementation, which improves over prior work on smooth pursuit text entry by 1) eliminating the need of calibration using motion correlation, 2) increasing input rate from 3.34 to 3.41 words per minute, 3) featuring text suggestions that were trained on 10,000 lexicon sentences recommended in the literature. We report on a user study (N=26) which shows that users are able to eye type at 3.41 words per minutes without calibration and without user training. Qualitative feedback also indicates that users positively perceive the system. Our work is of particular benefit for disabled users and for situations when voice and tactile input are not feasible (e.g., in noisy environments or when the hands are occupied).

In the recent years, intelligent vehicles became the most common everyday task that attracts a great interest from the industry. This work is dedicated to enhance the vehicles' intelligence through an adaptable four dimensional model capable of impersonating the vehicle driver by accurately predicting the time and location of destinations. Existing research efforts have addressed this challenge through prediction models that consider one or two aspects of human behavior, such as the social network or location semantics. However, the accuracy of these models remains rather limited. Accordingly, this paper addresses the shortfalls of the existing work by incorporating most aspects of human mobility. Moreover, we study the impact of each of the model dimensions on the prediction accuracy, individually and combined. We further propose an optimization algorithm to calculate the best blend of the model dimensions, that maximizes the prediction accuracy. The performance of the model is evaluated using a dataset that is collected from users in the city of UIm, Germany, as well as through computer simulations. The results show that the model can achieve a prediction accuracy of 95%, outperforming the state-of-the-art counterparts.

The Internet of Things (IoT) has enabled unprecedented interactions with our physical world, with the aim to deliver a wide range of customizable services in many domains. With recent advancements in IoT technology, users are increasingly expecting these services to be intelligent and context aware. Nevertheless, there is still no framework capable of delivering personalized IoT services on a large scale. For such a framework to be conceived, it is likely that technologies from many domains have to be utilized. This paper examines the readiness of the leading state-of-the-art technologies in several key fields for realizing the goal of a truly scalable and personalized IoT experience. We discuss the important requirements and challenges for realizing this goal. Then, we identify the major approaches that can contribute to this goal and categorize them into: technologies for adaptive personalized sensing, scalable solutions for user-centric networking, and intelligence techniques that leverage context awareness and adaptability at the application and system levels. In the first category, our discussion centers around virtualization and reprogrammability at the sensing layer. In the second category, we investigate the readiness of Fog computing and information-centric networking to develop scalable personalized IoT infrastructures. These approaches were chosen for their combined ability to match dynamic user requirements with available system resources, while guaranteeing overall efficient utilization. Finally, in the third category, we examine context awareness, reasoning, and machine learning techniques as well as semantic technologies for realizing proactive and adaptive intelligent IoT systems and applications. This paper offers a focused discussion of the key topics that drive the research in the important and timely topic of scalable and personalized IoT networks.

In the era of technology, Internet Of Things (IoT) which is a computing concept connecting a growing range of physical devices across the Internet, enriched our lives. IoT emerges with widespread applications and technologies including Wireless Sensor Networks (WSNs). Over the past few years, WSNs become one of the most evolving technologies and are considered as the connection between physical and virtual worlds for the ability to measure, collect and monitor environmental or physical conditions. Due to the large scale of WSNs being deployed to the Internet, security and privacy risks become crucial. The objective of this paper is to find a technique for anonymous communication and implement a security system for WSNs. This is achieved by implementing “Onion Routing” functionality on sensing devices, utilizing cryptography techniques and key distribution algorithms.

Blockchains enable mistrusting entities to agree on a common state, and preserve integrity without any need for a trusted 3rd party. However, blockchains suffer from scalability issues, where transaction throughput is affected by multiple factors such as inter-block time, number of peers, and blockchain forks. Accordingly, this paper proposes a novel decentralized peer-to-peer network layer that addresses these issues, and more. In our proposed Blockchain-Centric Exchange Protocol, an information-centric architecture is used to optimize information exchange. We demonstrate the ability of our design to meet many of today's application requirements such as permissioned blockchains, isolation of transaction, validation, and mining traffic, and eliminate the need for an overlay network. Using computer simulations we implement BCXP, compare its performance with other blockchain-centric protocols and IP overlay networks. The results show that even by pushing such IP overlay protocols to impractical limits, BCXP still provides clear performance advantages.

Most of the already existing authentication schemes are subject to multiple types of side-channel attacks such as shoulder surfing, smudge attacks, and thermal attacks. Meanwhile, motion sensors and eye trackers are becoming more accurate. We propose a novel authentication technique that leverages a combination of mid-air gestures and gaze input for shoulder surfing resilient authentication. The aim is to complicate shoulder surfing attacks by dividing the attacker's attention onto 1) the user's eyes, 2) hand-gestures, and 3) the screen. We report on the concept and implementation of the approach using both random and fixed layouts.

Gesture-based authentication systems are gaining increasing attention from the research community due to their promising usability. However, the scalability of these systems has not been properly investigated against the number of users and the number of gestures. Accordingly, in this paper, we explore the scalability of mid-air gesture-based systems in both aforementioned dimensions to enhance the already existing systems. We implemented a gesture-based authentication model with 20 gestures and we invited 39 users for data collection. A Support Vector Machine (SVM) classifier with Grid search cross-validation was used for training to prove the concept of the model's prototype. The obtained results proved that with the upscaling of the system from the aspect of the number of users, performance gets worse. On the other hand, as gestures introduced to the system increases, the performance improves.

As codebases for software projects get larger, reaching ranges of millions of lines of code, the need for computer-aided program comprehension grows. We define one of the tasks of program comprehension to be algorithm recognition: given a piece of source-code from a file, identify the algorithm this code is implementing, such as brute-force or dynamic programming. Most research in this area is making use of pattern matching, which involves much human effort and is of questionable accuracy when the structure and semantics of programs change. Thus, this paper proposes to let go of defined patterns, and make use of simpler features, such as counts of variables and counts of different constructs to recognize algorithms. We then feed these features to a classification algorithm to predict the class or type of algorithm used in this source code. We show through experimental results that our proposed method achieves a good improvement over baseline.

Road surface conditions can cause serious traffic accidents, often with tragic consequences. Thus, an efficient system for mapping road anomalies can significantly promote the safety of drivers and pedestrians. This paper proposes a novel road anomaly mapping system that is able to detect a wide variety of conditions with high accuracy. The smartphone's accelerometer and GPS sensors are used for detection to minimize infrastructure costs. In addition, to ensure the system is adaptive to different road conditions, pattern recognition techniques are used to automatically calculate the detection threshold. Furthermore, to compensate for GPS inaccuracies, reinforcement learning based on a proposed reward system is used to maximize confidence in the detected anomalies. The reward system is also able to forget anomalies that have been fixed. Moreover, the system is implemented in a distributed way between the smartphone and a cloud server to minimize cellular bandwidth usage, while still retaining the accuracy advantages of a centralized cloud. Live tests have been conducted to evaluate the performance of the system and the results show it is accurate under different driving conditions.

In a Public Safety (PS) situation, agents may require critical and personally identifiable information. Therefore, not only does context and location-aware information need to be available, but also the privacy of such information should be preserved. Existing solutions do not address such a problem in a PS environment. This paper proposes a framework in which anonymized Personal Information (PI) is accessible to authorized public safety agents under a PS circumstance. In particular, we propose a secure data storage structure along with privacy-preserving mobile search framework, suitable for Public Safety Networks (PSNs). As a result, availability and privacy of PI are achieved simultaneously. However, the design of such a framework encounters substantial challenges, including scalability, reliability of the data, computation and communication and storage efficiency, etc. We leverage Secure Indexing (SI) methods and modify Bloom Filters (BFs) to create a secure data storage structure to store encrypted meta-data. As a result, our construction enables secure and privacy-preserving multi-keyword search capability. In addition, our system scales very well, maintains availability of data, imposes minimum delay, and has affordable storage overhead. We provide extensive security analysis, simulation studies, and performance comparison with the state-of-the-art solutions to demonstrate the efficiency and effectiveness of the proposed approach. To the best of our knowledge, this work is the first to address such issues in the context of PSNs.

Driven by the advent of sophisticated and ubiquitous applications, and the ever-growing need for information, wireless networks are without a doubt steadily evolving into profoundly more complex and dynamic systems. The user demands are progressively rampant, while application requirements continue to expand in both range and diversity. Future wireless networks, therefore, must be equipped with the ability to handle numerous, albeit challenging, requirements. Network reconfiguration, considered as a prominent network paradigm, is envisioned to play a key role in leveraging future network performance and considerably advancing current user experiences. This paper presents a comprehensive overview of reconfigurable wireless networks and an in-depth analysis of reconfiguration at all layers of the protocol stack. Such networks characteristically possess the ability to reconfigure and adapt their hardware and software components and architectures, thus enabling flexible delivery of broad services, as well as sustaining robust operation under highly dynamic conditions. The paper offers a unifying framework for research in reconfigurable wireless networks. This should provide the reader with a holistic view of concepts, methods, and strategies in reconfigurable wireless networks. Focus is given to reconfigurable systems in relatively new and emerging research areas such as cognitive radio networks, cross-layer reconfiguration, and software-defined networks. In addition, modern networks have to be intelligent and capable of self-organization. Thus, this paper discusses the concept of network intelligence as a means to enable reconfiguration in highly complex and dynamic networks. Key processes in network intelligence, such as reasoning, learning, and context awareness, are presented to illustrate how these methods can take reconfiguration to a new level. Finally, the paper is supported with several examples and case studies showing the tremendous impact of reconfiguration on wireless networks.

Smart grids (SG) represent the next step in modernizing the current electric grid. In this structure, a communications network is combined with the power grid in order to gather information that can be used to increase the efficiency of the grid, reduce power consumption, and improve the reliability of services, among other numerous advantages. SG communication networks are unique in their large scale and the limited capabilities of nodes which present several challenges in the design of efficient routing protocols. This paper provides a comprehensive survey of the main networking challenges present in the design of SG communication networks, and some of the important routing protocols proposed to address those challenges. Various technologies and architectures proposed for routing in SGs are discussed. A detailed comparison of the protocols considered in this paper is also given, and key areas that require further investigation are highlighted.

In this paper, a novel cognitive engine for Wireless Sensor Networks (WSN) is proposed in order to achieve its end-to-end goals. This engine is designed using the tool known as Weighted Cognitive Maps (WCM). WCMs have the advantage of being able to consider multiple conflicting objectives and constraints with low complexity. Their inference properties also allow them to resolve complex network interactions using simple mathematical operations. Methods for designing the WCM system are illustrated. The performance of the proposed system is evaluated using computer simulations. Simulation results show that the WCM system outperforms its existing counterparts in metrics of network lifetime, throughput, and PLR.

Energy efficiency and network lifetime are key factors in characterizing wireless sensor networks due to the limited energy of nodes. In this paper we present two approaches to routing in wireless sensor networks that utilize the ideas of node cooperation and information exchange to achieve cognition across multiple network layers. In the first proposal, nodes exchange information about their statistical channel parameters to achieve awareness of the coverage area and use this information in path choice and transmit power adaptation. In the second proposal, nodes share information about energy states and utilize this information in achieving load balancing across nodes in the network. Nodes also cooperate with each other to reduce unnecessary transmissions. We evaluate our proposals through computer simulations and the results show that the energy efficiency of our proposals significantly outperform existing techniques, thus achieving the greater goal of extending network lifetime.