Fig 7 - uploaded by Sebastian von Mammen
Content may be subject to copyright.
Source publication
This paper establishes a connection between object tracking from a systems point of view and the job-scheduling or job-shop problem. Often, surveillance areas cannot be fully monitored by a set of smart cameras at any given point in time. Decisions have to be made, which objects are to be tracked. The computer vision aspects of object tracking have...
Context in source publication
Context 1
... camera which executes pan- tilt-functionality infinitely fast. As a result, the camera would monitor its environment infinitely fast, yielding a closed-ring panoramic image. A monitoring plan would not be necessary in this case. For this reason, the number and location of all persons within a specific distance (r F oV ) to a camera are known. In Fig. 7 an according scenario is illustrated. ...
Similar publications
Person re-identification (Re-ID) is the task of matching humans across cameras with non-overlapping views that has important applications in visual surveillance. Like other computer vision tasks, this task has gained much with the utilization of deep learning methods. However, existing solutions based on deep learning are usually trained and tested...
![Figure 3: Example images of German train station cameras. Source: [5]](profile/Alen-Smajic-2/publication/344830620/figure/fig23/AS:950542201257984@1603638329132/Example-images-of-German-train-station-cameras-Source-5_Q320.jpg)
![Figure 3: Example images of German train station cameras. Source: [5]](profile/Alen-Smajic-2/publication/344830620/figure/fig24/AS:950542226452483@1603638335785/Example-images-of-German-train-station-cameras-Source-5_Q320.jpg)
![Figure 3: Example images of German train station cameras. Source: [5]](profile/Alen-Smajic-2/publication/344830620/figure/fig25/AS:950542247403521@1603638340554/Example-images-of-German-train-station-cameras-Source-5_Q320.jpg)
Recently there was a nationwide scandal about an incident at Frankfurt Central Station in which a boy was pushed onto the railroad tracks in front of an arriving train and lost his life due to its injuries. However, this incident is not the only one of its kind because such accidents occur from time to time at train stations. At the moment there ar...
Citations
... The Self-Organizing Smart Camera System [22] [28] consists of a number of cameras with computation power. They communicate and work together to autonomously track persons or groups of persons. ...
The growing complexity of current computer systems requires a high amount of administration, which poses an increasingly challenging task for manual administration. The Autonomic and Organic Computing Initiatives have introduced so called self-x properties, including self-configuration, self-optimization, self-healing, and self-protection, to allow administration to become autonomous. Although research in this area revealed promising results, it expects all participants to further the system goal, i.e., their benevolence is assumed. In open systems, where arbitrary participants can join the systems, this benevolence assumption must be dropped, since such a participant may act maliciously and try to exploit the system. This introduces a not yet considered uncertainty, which needs to be addressed.
In human society, trust relations are used to lower the uncertainty of transactions with unknown interaction partners. Trust is based on past experiences with someone, as well as recommendations of trusted third parties. In this work trust metrics for direct trust, reputation, confidence, and an aggregation of them are presented. While the presented metrics were primarily designed to improve the self-x properties of OC systems they can also be used by applications in Multi-Agent-Systems to evaluate the behavior of other agents. Direct trust is calculated by the Delayed-Ack metric, that assesses the reliability of nodes in Organic Computing systems. The other metrics are general enough to be used with all kinds of contexts and facets to cover any kind of trust requirements of a system, as long as corresponding direct trust values exist. These metrics include reputation (Neighbor-Trust), confidence, and an aggregation of them.
Evaluations based on an Automated Design Space Exploration are conducted to find the best configurations for each metric, especially to identify the importance of direct trust, reputation, and confidence for the total trust value. They illustrate, that reputation, i.e., the recommendations of others, is an important aspect to evaluate the trustworthiness of an interaction partner. In addition, it is shown that a gradual change of priority from reputation to direct trust is preferable instead of a sudden switch when enough confidence in the correctness of ones own experiences is accumulated. All evaluations focus on systems with volatile behavior, i.e., system participants change their behavior over time. In such a system, the ability to adapt fast to behavior changes has turned out to be the most important parameter.
Augmented virtual environments (AVE) combine real-time videos with 3D scenes in a Digital Earth System or 3D GIS to present dynamic information and a virtual scene simultaneously. AVE can provide solutions for continuous tracking of moving objects, camera scheduling, and path planning in the real world. This paper proposes a novel approach for 3D path prediction of moving objects in a video-augmented indoor virtual environment. The study includes 3D motion analysis of moving objects, multi-path prediction, hierarchical visualization, and path-based multi-camera scheduling. The results show that these methods can give a closed-loop process of 3D path prediction and continuous tracking of moving objects in an AVE. The path analysis algorithms proved accurate and time-efficient, costing less than 1.3 ms to get the optimal path. The experiment ran a 3D scene containing 295,000 triangles at around 35 frames per second on a laptop with 1 GB of graphics card memory, which means the performance of the proposed methods is good enough to maintain high rendering efficiency for a video-augmented indoor virtual scene.
This research considers the problem of maximizing information collection and exchange between unmanned resources and a control station in a bandwidth constrained environment. Unmanned aerial vehicles (UAVs) are utilized in disaster response to gather data and aid in intelligence, surveillance, and reconnaissance (ISR). They are regularly equipped with multiple gimbal-mounted heterogeneous sensors, generating large amounts of data which are to be routed through a communications network back to mission managers. A mixed integer linear program (MILP) is formulated for this problem. However, due to the complexity of this problem, the MILP is only able to solve trivially-sized problems. A three-phase heuristic approach is proposed which, through extensive testing, is proven to efficiently solve large-sized problems. Our analysis also gives insight into the features of this problem and their impact on mission success.
Open self-x systems of a very large scale – interconnecting several thousand of autonomous and heterogeneous entities – become increasingly complex in their organisational structures. This is due to the fact that such systems are typically restricted to a local view in the sense that they have no global instance, which can be responsible for controlling or managing the whole system. Therefore, new ways have to be found to develop and manage them. An essential aspect that has recently gained much attention in this kind of systems is the social concept of trust. Using appropriate trust mechanisms, entities in the system can have a clue about which entities to cooperate with. This is very important to improve the robustness of self-x systems, which depends on a cooperation of autonomous entities. The contributions of this chapter are trustworthy concepts and generic self-x algorithms with the ability to self-configure, self-optimise, and self-heal that work in a distributed manner and with no central control to ensure robustness. Some experimental results of our algorithms are reported to show the improvement that can be obtained compared with the baseline measurements.
This paper presents a holistic approach to execute tasks in distributed smart systems. This is shown by the example of monitoring tasks in smart camera networks. The proposed approach is general and thus not limited to a specific scenario. A job-resource model is introduced to describe the smart system and the tasks, with as much order as necessary and as few rules as possible. Based on that model, a local algorithm is presented, which is developed to achieve optimization transparency. This means that the optimization on system-wide criteria will not be visible to the participants. To a task, the system-wide optimization is a virtual local single-step optimization. The algorithm is based on proactive quotation broadcasting to the local neighborhood. Additionally, it allows the parallel execution of tasks on resources and includes the optimization of multiple-task-to-resource assignments.
Distributed sensing networks are getting increasingly complex these days. The main reason are the changing demands of the users and application scenarios, which require multipurpose systems. Enabled by continuously improving computational and storage capacities of sensors, this development leads to an increasing number of different algorithms which run concurrently in a sensing network. Thereby, they enable sensor-actuator platforms to perform various kinds of analysis and actions in parallel. Within such a sensor network a variety of algorithms is performed simultaneously. When developing distributed vision and control algorithms, developers focus mainly on the consecutive processing stages. Such a process typically begins with perceiving raw sensor data and terminates with delivering high-level event data to responsible entities. Thereby, different stages may be performed at varying locations within the underlying network. Although the researchers may apply custom optimizations to their data flows, these are highly specific. During design time, it is impossible to anticipate each system environment or predict their algorithms’ possible interactions and synergies with other data flows. We propose a generic storage architecture which separates algorithms from data storage and retrieval. By making use of the fact that most data in sensing networks refers to geographic areas, our architecture takes care of the data flow and its online optimization throughout the network at runtime. By decoupling the processing stages from the data flow, we allow for self-organizing meta-level optimizations of data placement in the network. Moreover, this approach even makes inter-algorithmic optimizations possible, if different algorithms process similar data within their step-wise processing logic. With the introduction of the access-centric storage paradigm, we prove to reduce network load and query latency at the same time at runtime.
