Premysl Volf

Agent-Based Multi-Layer Collision Avoidance to Unmanned Aerial Vehicles

This contribution presents a distributed, multi-layer collision avoidance architecture supporting efficient utilization of air space shared by several autonomous aerial vehicles. Presented multi-layer architecture is based on deliberative deployment of several collision avoidance methods by the aircraft at the same time. Both cooperative and non-cooperative collision avoidance methods are presented in the paper. The robustness of the architecture is justified by means of experimental validation of multi-agent simulation

Multi-party Collision Avoidance among Unmanned Aerial Vehicles

This paper addresses the problem of distributed cooperative collision avoidance that supports efficient utilization of air space shared by several autonomous unmanned aerial vehicles. The novel multi-party collision avoidance (MPCA) algorithm is described. It is compared to the iterative peer-to-peer collision avoidance (IPPCA) algorithm that iteratively optimizes social welfare. The paper provides a set of experiments and a comparison of different collision avoidance mechanisms in a multi-agent model of air traffic.This paper addresses the problem of distributed cooperative collision avoidance that supports efficient utilization of air space shared by several autonomous unmanned aerial vehicles. The novel multi-party collision avoidance (MPCA) algorithm is described. It is compared to the iterative peer-to-peer collision avoidance (IPPCA) algorithm that iteratively optimizes social welfare. The paper provides a set of experiments and a comparison of different collision avoidance mechanisms in a multi-agent model of air traffic.

Convergence of Peer-to-Peer Collision Avoidance among Unmanned Aerial Vehicles

In this article we study the theoretical aspects of the collision avoidance among the collectives of unmanned aerial vehicles (UAVs) engaged in the free flight operation. The free flight based operations do not provide collision-free trajectories as a flight specification to the UAVs. Collision avoidance of the trajectories is implemented by means of peer-to-peer negotiations among the individual UAVs. In this paper we study the role of interaction in the collision avoidance process. We prove theoretically convergence of the specific negotiating protocol that has been deployed in a practical implementation of the software prototype of the distributed collision avoidance system.In this article we study the theoretical aspects of the collision avoidance among the collectives of unmanned aerial vehicles (UAVs) engaged in the free flight operation. The free flight based operations do not provide collision-free trajectories as a flight specification to the UAVs. Collision avoidance of the trajectories is implemented by means of peer-to-peer negotiations among the individual UAVs. In this paper we study the role of interaction in the collision avoidance process. We prove theoretically convergence of the specific negotiating protocol that has been deployed in a practical implementation of the software prototype of the distributed collision avoidance system.

LARGE-SCALE HIGH-FIDELITY AGENT-BASED SIMULATION IN AIR TRAFFIC DOMAIN

We present the concept of dynamic partitioning of scalable, high-fidelity multi-agent simulation complemented with intelligent load-balancing processes. The simulation framework is designed to simulate entities to high details that require extended computation resources. To be able to simulate a huge amount of entities, distributed simulation is introduced using spatial partitioning and dynamic load balancing. A novel and important feature is the combination of the synchronous and asynchronous parts in the simulation. We use the domain of the air traffic simulation to verify the simulation framework. We present a method to perform spatial and temporal planning within 3D space and multilayer architecture using several collision avoidance algorithms to illustrate the high computational demands of each airplane. The platform has been used to support simulation of an entire civilian air traffic touching the national airspace of United States. A thorough evaluation of the system has been performed, confirming that it can scale up to a very high number of complex agents operating simultaneously (thousands of aircrafts) with full detailed models.

RBVO Formation Protocol

The proposed protocol has been designed to support a flexible formation of request-based virtual organizations with an emphasis on reflecting the conditions of real competitive environments. It supports automated or semi-automated negotiations mainly in the creation part of a virtual organization life cycle and accounts for a use of service level agreements. The protocol consists of three phases: (i) potential partner search, (ii) negotiation of SLAs and RBVO establishment, and (Hi) RBVO execution and dissolution. The protocol is based on FIPA standards.

Agentfly: NAS-wide simulation framework integrating algorithms for automated collision avoidance

AgentFly is a software prototype providing a distributed architecture for large-scale NAS-wide simulation implemented as a multi-agent system. AgentFly is implemented on top of the Aglobe [1] platform which is both an implementation framework and a runtime engine for custom agents. It was selected over possible alternatives (e.g. JADE [2]) for its outstanding performance and scalability supporting seamless interaction among heterogeneous software, hardware and human actors. AgentFly system has been developed for over five years. It was initially built f or simulation-based validation and comparison of various approaches for autonomous collision avoidance algorithms adopting the free-flight concept. Later, AgentFly has been extended with high-level control algorithms providing tactical control — i.e. coordination of several autonomous unmanned aerial vehicles (UAV). The same agents and algorithms integrated in AgentFly simulation are also deployed on real UAV platforms. Besides this UAV-related application, the U.S. Federal Aviation Administration (FAA) supports the application of the AgentFly system for simulation and evaluation of the future civilian air-traffic management system. AgentFly has been extended with high-fidelity computational models of civilian airplanes and a parallelization concept integrating dynamic load-balancing. The parallelized approach of AgentFly has been validated in simulation using data of a full civilian air-traffic touching NAS. Nowadays, AgentFly is being extended so that it provides a simulation of ATC functions for the NEXTGEN concept validation. There are being integrated ATC & NAS automation agents which are used to simulate human operation in ATM.

Distributed Planning Algorithm for Coalition Logistics in Semi-trusted Environment

We present a collective approach to coalition logistics planning that presents the features crucial for application in an adversarial environment: planning and communication efficiency, well-defined levels of information to be shared, tight integration of trustfulness with the planning and stability with respect to imprecise trustfulness values. To achieve this goal, we combine multi-agent negotiation with efficient fuzzy and flexible linear programming techniques from operation research field. Alternating rounds of global optimization and restricted negotiation split the task into sub-tasks, create teams, assign them to the tasks and provide a task-resource mapping. Resulting plan execution can be easily verified and verification results can be used to update the trust and social models and potentially to perform re-planning immediately

Airspace sectorization optimization using fast-time simulation of air traffic controller's workload

The air-traffic management (ATM) is a very complex system. Increasing air traffic and demands on efficiency of the ATM system requires improvement of the ATM system. One of possible improvements is to optimize sectorization of the airspace to improve selected metrics. Each proposed sectorization has to be evaluated to determine impact of changes to current system. Evaluation usually consists of several approaches to cover all aspects of the change - e.g. expert groups, statistical data, real-time simulations, fast-time simulations, etc. The paper presents fast-time simulation system AgentFly used to evaluate proposed variants of sectorization. The simulation is able to compute and evaluate metrics important for a decision making of selected sectorization. This approach is applied to a sectorization of the airspace of the Czech Republic. Scenarios were simulated using current and predicted future air traffic. The AgentFly gathers data related to aircraft and ATC to compute requested metrics. Simulation outputs were used for evaluation and comparison of proposed sectorizations.

Simulation of UAS integration into shared airspace for validation of impact on ATM systems

In this paper we introduce a modular modeling and simulation framework, which allows validating various aspects of UAS integration into shared airspace and its impact on existing ATM systems. Framework supports large-scale fast-time simulations, as well as real-time simulations with various virtualization levels, i.e. individual software components might be replaced by hardware-in-the-loop or human-in-the-loop to increase simulation fidelity. This can simplify and speed up the process of deployment on new concepts addressing various issues of UAS integration in different airspace classes.

Optimization-Based Collision Avoidance for Cooperating Airplanes

Application of autonomous intelligent systems into airspace domain is very important nowadays. The paper presents decentralized collision avoidance algorithm utilizing a solution of the defined optimization problem where efficiency criteria, collision penalties and airplanes’ missions are integrated in an objective function. Two different implementation approaches used for stochastic Probability Collectives optimizer are presented and evaluated – a complex distributed multi-agent deployment among participating airplanes and the Process Integrated Mechanism inspired architecture. Both approaches have been validated and evaluated on the multi-agent framework AGENTFLY providing precise simulation for airspace operations.