Swedish Institute of Computer Science Publications Database
Not a member yet
2787 research outputs found
Sort by
An Experimental Study of Attacks on the Availability of Glossy
No full textGlossy is a reliable and low latency flooding mechanism designed primarily for distributed communication in wireless sensor networks (WSN). Glossy achieves its superior performance over tree-based wireless sensor networks by exploiting identical concurrent transmissions. WSNs are subject to wireless attacks aimed to disrupt the legitimate network operations. Real-world deployments require security and the current Glossy implementation has no built-in security mechanisms. In this paper, we explore the effectiveness of several attacks that attempt to break constructive interference in Glossy. Our results show that Glossy is quite robust to approaches where attackers do not respect the timing constraints necessary to create constructive interference. Changing the packet content, however, has a severe effect on the packet reception rate that is even more detrimental than other physical layer denial-of-service attacks such as jamming. We also discuss potential countermeasures to address these security threats and vulnerabilities
Production planning for district heating networks (DHN)
No full textDistrict Heating Networks (DHN) can provide higher efficiencies and better pollution control
compared to local heat generation. However, there are still many areas, which can be improved and
optimized in these systems. A DHN is a complex distributed system of different customer substations
and components such as boilers, accumulators, pipes, and in many cases also turbines for
electricity production. How to schedule the components with the objective of maximizing the profit
of heat and electricity production over a finite time horizon is receiving increased attention, and
is the problem that has been dealt with in this work. This mixed integer linear programming (MILP)
problem has been formulated as a unit commitment problem (UCP), which involves finding the most
profitable unit dispatch regarding production costs and heat and electricity sell prices, while
simultaneously meeting the predicted district heating demands and satisfying network operational
constraints. The heating demands within the optimization time horizon are predicted based on season
and weather forecasts.
In this work, the district heating plant in Uppsala, Sweden, owned by Vattenfall AB, has been
considered as a reference plant for modeling and optimization. The optimization model is
formulated in Python using Pyomo modeling language, and solved by Gurobi and GLPK solvers. An
hourly- based data of five consecutive days is used as the time horizon. The results demonstrate
the fact that with an accurate model of the DHN, it is possible to significantly increase the
revenue of the DHN by finding the most economical way to dispatch different production components
Demo Abstract: SicsthSense - Dispersing the Cloud
No full text—This demo presents SicsthSense, our open cloud
platform for the Internet of Things. SicsthSense enables low
power devices such as sensor nodes and smartphones to easily
store their generated data streams in the cloud. This allows the
data streams, and their history, to be made permanently available
to users for visualisation, processing and sharing. Moving sensor
data computation and monitoring into the cloud is a promising
avenue to enable centralisation of control and redistribution of
collected data.
We showcase SicsthSense running with real sensor nodes
collecting environmental data and posting it to our datastore.
This live data is then visualised and made available for sharing
between users of the platform. Our Android App will also be
distributed to enable participants to stream their phone sensors
into the system, demonstrating how simple it can be to start
machine-to-machine interactions with SicsthSense
Large Scale Characterisation of YouTube Requests in a Cellular Network
No full textTraffic from wireless and mobile devices is expected
to soon exceed traffic from fixed devices. Understanding the
behaviour of users on mobile devices is important in order to
improve the offered services and the provision of the underlying
network. Globally, more than 60% of consumer Internet traffic is
estimated to be video traffic, and the most popular video website,
YouTube, estimates that mobile access makes up nearly 40% of
the global watch time. This paper presents the first work to study
the characteristics of YouTube user requests on a nationwide
cellular network. This study is based on the analysis of a large
dataset generated by 3 million users and collected by a major
telecom operator. We show for instance that 20% of the users
generate 78% of the requests, and that over 80% of the requests
target only 20% of the distinct videos accessed during the data
collection period. Our results provide a comprehensive insight
into the way people use YouTube on mobile devices, and show a
very high potential for video cacheability on the cellular network
Case-Based Reasoning for Explaining Probabilistic Machine Learning
No full textThis paper describes a generic framework for explaining the prediction of probabilistic machine learning algorithms using cases. The framework consists of two components: a similarity metric between cases that is defined relative to a probability model and an novel case-based approach to justifying the probabilistic prediction by estimating the prediction error using case-based reasoning. As basis for deriving similarity metrics, we define similarity in terms of the principle of interchangeability that two cases are considered similar or identical if two probability distributions, derived from excluding either one or the other case in the case base, are identical. Lastly, we show the applicability of the proposed approach by deriving a metric for linear regression, and apply the proposed approach for explaining predictions of the energy performance of households
Characteristics of Software Ecosystems for Federated Embedded Systems: A Case Study
No full textContext:
Traditionally, Embedded Systems (ES) are tightly linked to physical products, and closed both for communication to the surrounding world and to additions or modifications by third parties. New technical solutions are however emerging that allow addition of plug-in software, as well as external communication for both software installation and data exchange. These mechanisms in combination will allow for the construction of Federated Embedded Systems (FES). Expected benefits include the possibility of third-party actors developing add-on functionality; a shorter time to market for new functions; and the ability to upgrade existing products in the field. This will however require not only new technical solutions, but also a transformation of the software ecosystems for ES.
Objective:
This paper aims at providing an initial characterization of the mechanisms that need to be present to make a FES ecosystem successful. This includes identification of the actors, the possible business models, the effects on product development processes, methods and tools, as well as on the product architecture.
Method:
The research was carried out as an explorative case study based on interviews with 15 senior staff members at 9 companies related to ES that represent different roles in a future ecosystem for FES. The interview data was analyzed and the findings were mapped according to the Business Model Canvas (BMC).
Results:
The findings from the study describe the main characteristics of a FES ecosystem, and identify the challenges for future research and practice.
Conclusions:
The case study indicates that new actors exist in the FES ecosystem compared to a traditional supply chain, and that their roles and relations are redefined. The business models include new revenue streams and services, but also create the need for trade-offs between, e.g., openness and dependability in the architecture, as well as new ways of working
All is not Lost: Understanding and Exploiting Packet Corruption in Outdoor Sensor Networks
No full textDuring phases of transient connectivity, sensor nodes receive a substantial number of corrupt packets. These corrupt packets are generally discarded, losing the sent information and wasting the energy put into transmitting and receiving. Our analysis of one year’s data from an outdoor sensor network deployment shows that packet corruption follows a distinct pattern that is observed on all links. We explain the pattern’s core features by considering implementation aspects of low-cost 802.15.4 transceivers and independent transmission errors. Based on the insight into the corruption pattern, we propose a probabilistic approach to re- cover information about the original content of a corrupt packet. Our approach vastly reduces the uncertainty about the original content, as measured by a manifold reduction in entropy. We conclude that the practice of discarding all corrupt packets in an outdoor sensor network may be unnecessarily wasteful, given that a considerable amount of information can be extracted from them
Parallel Community Detection For Cross-Document Coreference
No full textThis document presents a highly parallel solution for cross-document coreference resolution, which can deal with billions of documents that exist in the current web. At the core of our solution lies a novel algorithm for community detection in large scale graphs. We operate on graphs which we construct by representing documents' keywords as nodes and the co-location of those keywords in a document as edges. We then exploit the particular nature of such graphs where coreferent words are topologically clustered and can be efficiently discovered by our community detection algorithm. The accuracy of our technique is considerably higher than that of the state of the art, while the convergence time is by far shorter. In particular, we increase the accuracy for a baseline dataset by more than 15\% compared to the best reported result so far. Moreover, we outperform the best reported result for a dataset provided for the Word Sense Induction task in SemEval 2010
Designing auditory alarms for an industrial control room
No full textConstantly increasing and more complex information flows in industrial control rooms raise the risk that operators will become distracted, confused and visually overloaded in demanding situations. New multimodal interfaces may offer better solutions. In this work we focus on alarm sound design. Alarms serve to alert operators to deviations from
normal conditions and enable them to react appropriately. The speed and accuracy with which operators can identify alarms are crucial to effectiveness. Additionally, auditory alarms should not be too annoying or distracting. Salient auditory stimuli effectively catch and guide attention, regardless of operators’ visual focus. Research has established that auditory signals can be designed to express
different levels of urgency. Sound can also convey detailed information. However auditory alarms are often implemented carelessly, using sounds that are too loud, too many and too confusing.The aim of this work was to develop a concept to enhance the auditory alarms in a control room of a paper mill. The goals were to improve operator effectiveness and acceptance. Before concept development a pretest involving 21 operators evaluated the state of the alarm sounds. The results indicated poor design and confirmed some well-known issues. The developed concept consists of new alarm sounds, spatial presentation of the sounds and alarm repetition intervals. The sounds convey urgency information and information associated with the production sections. The user-centred design process involved 24 operators and ontained iterations in which the concept was refined and feedback incorporated. An evaluation involving operators then studied the effects of the new concept. Operator acceptance was assessed using the Van der Laan acceptance scale, rating the usefulness and satisfaction of alarm sounds. There was a considerable increase in both usefulness and satisfaction scores between the pretest and posttest. These results support that the developed concept increases operator effectiveness and acceptance