Claudius Stern

A distributed multi-UAV path planning approach for 3D environments

During rescue scenarios it is indispensable to obtain an overview of the situation. Unmanned aerial vehicles (UAVs) can gather the necessary information in a fast and efficient way. This paper presents an approach for path planning in 3D environments offering a solution to explore disaster areas including, e. g., partially or completely destroyed buildings. Using multiple UAVs decreases the time needed to receive a complete overview if the problem of coordination and task allocation is solved. We present an approach for the use of multiple UAVs. The UAVs work in a distributed manner without any central coordination instance and cover the exploration of terrains as well as goal-oriented path planning. When using multiple UAVs redundant exploration is avoided through the use of inter-UAV-communication. The approach is based on potential fields and uses the simplicity of the gradient method to calculate paths for fast exploration of the terrain.

Combining Autonomous Exploration, Goal-Oriented Coordination and Task Allocation in Multi-UAV Scenarios

Successful rescue operations after big accidents or natural disasters require a fast and efficient overview of the overall situation. With recent advances, unmanned aerial vehicles (UAVs) are more and more a viable choice under such circumstances.With the number of employed UAVs, the problem of coordination arises as well as proper task allocation among possibly heterogeneous UAVs. This paper presents a hybrid approach for UAV coordination and covers the exploration of unknown terrains as well as goal-oriented coordination and simultaneous task allocation. The approach combines the simplicity of the gradient method with informed A* search and supports prioritized task assignment. The system is suited for highly dynamic environments requiring frequent path recalculations.

Increasing motion information by using universal tracking of 2D-features

As long as visual features for recognition are known in advance and remain static due to a controlled environment, object tracking is state-of-the-art. Tracking objects in dynamically changing environments is still a challenge. Even harder is the tracking of moving objects with a moving camera. Our algorithm realizes a deterministic approach to track any 2D-features representable in a general way. Using our algorithm results in a more dense motion flow field and objects that move with a different motion vector than the environment can be clearly identified.

A 3D Path Planning Approach Extended by Bifurcation Theory for Formation Flights

After big disasters, search and rescue missions take place. During these missions it is essential to obtain an overview of the overall situation to arrange efficient rescue tasks as soon as possible. Afterwards, this information has to be continuously updated during the complete mission in order to quickly respond to changing conditions. The use of unmanned aerial vehicles (UAVs) is a viable choice to obtain such an overview in a fast and efficient way. Using multiple UAVs, the problem of coordination has to be solved to decrease the time needed to explore an area. We present an approach for the coordination of UAVs by setting up formation patterns using bifurcation theory. We combine this theory with a potential field approach for exploration, based on harmonic functions.

Role-based path planning and task allocation with exploration tradeoff for UAVs

After large accidents or natural disasters a fast and efficient retrieval of injured is required. With recent advances, unmanned aerial vehicles (UAVs) are more and more a viable choice under such circumstances. The problem of coordination arises with the number of employed UAVs as well as proper task allocation among possibly heterogeneous UAVs. In this work a hybrid approach for UAV coordination and task allocation is presented. It covers the exploration of unknown terrains as well as role-based, goal-oriented path planning and simultaneous task allocation. The approach uses a role system to support prioritized task assignment and an informed search algorithm to find consistent trajectories to known goal positions. The system is suited for highly dynamic environments, requiring frequent recalculations.

Dynamic optimization of track sectioning in mechatronic systems for a hybrid planner

For railbound autonomous systems a hybrid planner chooses a path along track sections. The vehicles propulsion uses a linear motor that needs trajectories for an air gap adjustment system. To model these, physical and fuzzy models as well as historical data of the air gap are used to predict the course of the air gap. The planner uses simulation to evaluate actions that have complete track sections as a granularity basis. The plan quality is improved by optimizing the choice of track sections after an analysis of the perceived and expected air gap gradient as that constitutes a major part of the systems overall performance.

Towards using virtual forces for image registration

This paper introduces a novel method for registering a large amount of aerial images when camera parameters are almost unknown and no reference images are available. The envisioned application is the creation of an overview map of a disaster area from images made by unmanned aerial vehicles (UAVs) equipped with cameras. The camera systems only have uncertain information about flight attitude. With traditional methods, the relatively small perspective errors per image sum up over time resulting in perspective errors which prevent building a consistent map from successively arriving images. By using virtual forces between images, the image parameters are continuously adapted to the current map. The small projective errors of quasi orthographic images are distributed among overlapping images. Having UAVs which deliver quasi orthographic images, our approach can build a contemporary overview without the need to have all images in advance.

Integrating sporadic imitation in Reinforcement Learning robots

Although the combination of Reinforcement Learning and imitation has been already considered in recent research, it always revolved around fixed settings where demonstrator and imitator are fixed and the imitation process is a well-defined period of time. What is missing is the investigation of approaches that also work in scenarios where imitation is only sporadically possible. This means that in a multi-robot scenario a robot is now allowed to interrupt another robot by asking to repeat certain actions, but can only observe and integrate information bits delivered occasionally. In this paper we present how that can be done in continuous and noisy environment within an SMDP context.

Reliable Evidence of Data Integrity from an Untrusted Storage Service

Digital storage demand is growing with the increasing use of digital artifacts from media files to business documents. Regulatory frameworks ask for unaltered, durable storage of business communications. In this paper we consider the problem of getting reliable evidence of the integrity and existence of some data from a storage service even if the data is not available for reference anymore and even if the partner may be untrustworthy or malicious. Several solutions are presented, among them the first cryptographically secure approach providing evidence of the integrity of remotely stored data, without owning or transferring the complete original data.

Hierarchically Distributing Embedded Systems for Improved Autonomy

Distribution of functionality among nodes is a contemporary research issue for embedded systems, e.g. in the field of autonomous mobile robot groups. In such groups, the concept of distribution is mainly used to achieve flexibility and robustness that could not be reached by a single robot. Here it will be used as a design-paradigm for a robot’s internal architecture. In this paper, a hierarchically distributed robot architecture will be introduced which leads to an improved autonomy of the overall system.