Euroswarm

Unmanned Heterogeneous Swarm of Sensor Platforms



UNMANNED HETEROGENEOUS SWARM OF SENSOR PLATFORMS

Autonomous Systems (AS), which unmanned swarm systems are part of, will make a significant and revolutionary social, economic, education and research impact immediately. Use of swarm technologies and techniques has the potential to represent more than an evolution for the military doctrine and for the implementation of military missions: they could extend the reach and access of operations, reduce risk to warfighters, provide increased capability across the battlespace. In the longer term, swarm technologies can significantly reduce the cost of acquisition and operations of defence systems while minimizing human risk. EuroSWARM has the ambitious goal to become the benchmark in the unmanned heterogeneous swarm systems for defence applications. The EuroSWARM project objectives are:

  1. Develop the following key techniques for adaptive, informative, and reconfigurable operations of unmanned heterogeneous swarm systems: (a) Optimal task allocation and resource management (b) Sensor fusion (c) Cooperative guidance (d) Robust sensor network;
  2. Integrate the developed techniques;
  3. Validate the developed techniques based on simulations targeting specific military scenario;
  4. Demonstrate the proposed solutions based on a small-scale demonstrator in both indoor and outdoor environments.

The main output of the project will result in a modular, scalable and flexible swarm architecture which, in combination with a low-cost demonstration based on COTS devices, will represent the first step for the progressive uptake of unmanned swarm technology and applications in the defence sector. The EuroSWARM autonomous swarm system of heterogeneous sensor, can become a pilot for large scale implementation of such technology for critical European and Global challenges such as border control, surveillance-security, and with a clear dual-use potential.


Applications

Swarm based autonomous systems can be complex and most systems are ‘customized’ for specific scenarios. In EuroSWARM a modular, flexible and robust architecture is proposed using COTS sensors, innovative GNC algorithms along with a low cost practical demonstration with miniature unmanned aerial and ground vehicles for various task or mission scenarios.
However, other task scenarios, such as potential target proposals and combat search and rescue, will be also considered in developing the enabling technologies. One of the main considerations in the development will be that the proposed technologies should be able to be naturally evolved and easily extended to other task scenarios.
Systems to be considered in the unmanned swarm system consist of two main types: static/unattended sensor network and mobile sensor platforms.
The key enabling technologies identified to realize the concept of the proposed research project are:

  1. Optimal sensor network design of the static/unattended sensor network (resource allocation)
  2. Dynamic task allocation of the mobile sensor platforms
  3. Frequency and bandwidth assignment of the sensor network
  4. Sensor fusion
  5. Information fusion: behaviour monitoring
  6. Cooperative FDIR

These technologies are closely related and an overview of their inter-relation is illustrated in Figure 1. In this project, all these enabling technologies will be developed sharing a common objective, which is maximizing the situational awareness information.

Mission

Mission Scenario

A mission scenario with a very strong interest by military and security/law enforcement agencies is the case in which a specific area of high interest requires persistent monitoring/surveillance. It is assumed that the scenario takes place at the battlefield in conflict with a well-armed and competent opponent. At the moment, the battle is not open but the situation is tense, skirmishes occur frequently, and the risk for sabotage is considerable. Conflict escalation must be avoided, at the same time as own troupes and assets need to be effectively protected. A high value asset (HVA) is considered in this mission scenario as a strategically important camp with fuel deposits and stashes of ammunition and other military supplies. A company is allocated for camp protection and daily maintenance. The camp is located in rough and partly hilly terrain, but it is assumed that the existence and importance of the HVA is known to enemy formations in the region. Guards relieve each other in the task of surveying the camp perimeter for intrusions and to early identify sabotage attempts and enemy intelligence missions.


The sensor system

The guards use a sophisticated sensor system to support the perimeter surveillance. As the terrain limits the visibility in the protected area, centralized sensors are however ineffective. Instead the guards use ground sensors distributed in a large area around the camp that facilitates early indications on enemy reconnaissance or approaching formations. The ground sensors are sensitive for the presence of humans, vehicles and animals and give prompt alarms if potential targets are in the vicinity. The ground sensors have limited capability to assess the nature and severity of the threats, but are on the other hand robust and persistent and therefore they enable large area coverage for an attractive price of purchase and effort of maintenance. To support alarm verification, the network of ground sensors is completed with drones that operate autonomously in the sensor system. The drones are only activated for special tasks and missions. For instance, on a ground sensor alarm, the drones do a verification that may give the guard high fidelity video and other sensor data from the target or target area in real time. The drones are autonomous and can collaborate in pair, group or swarm to meet high demands on service quality and persistence toward hostile means of deception and attacks.

Operational Requirements

  1. Use a swarm of heterogenous vehicles (5-10)
    • Innovation: first full integration and testing of 10 heterogenous vehicles (static sensors, fixed wing UAV, quadrotors, UGVs) fully autonomously
    • Limitations: small size of drones, limited bandwidth, time, effort
  2. Integration and autonomous operation of a swarm of UxVs
    • Innovation: full integration of core technology blocks: sensor fusion, tasking, target tracking, UxV autonomy
    • Limitations: small testing area, medium level of autonomy
  3. Application of a swarm of UxVs in a realistic, representative military mission scenario
    • Innovation: selected the persistent monitoring scenario of a high value asset
    • Limitations: experiment time, asset complexity/mobility


Functional Requirements

  1. Sensor network
    • Maximisation of the sensor coverage
    • Maximising target detection
  2. Sensor fusion
    • Target detection, identification, and tracking
  3. Information fusion: behaviour monitoring
    • Detection of suspicious behaviours exhibited by the target
    • Identification of suspicious behaviours
  4. Mobile tasking
    • Task assignment for the mobile sensing platforms
    • Helping to maximise the situational awareness information
  5. Cooperative guidance
    • Maintaining persistent sensing
    • Maintaining the target tracking


Technology

Project Objectives and Challenges

Objectives

  1. Develop novel techniques for adaptive, scalable and robust operations of unmanned heterogeneous swarm
  2. Integrate and validate the developed technology using a realistic high value asset protection military scenario
  3. Demonstrate the proposed solutions based on a small scale set of experiments


Challenges

  1. Setup of the architecture and network
    • Different autopilots with MAVLink communication protocols, various hardware
  2. Flight endurance
    • Low battery capacity
  3. Sensor errors
    • Relatively big noise sources in low cost, small size and light weighted sensors
    • The noise sources propagate through the swarm key functions and could significantly degrade their performance
  4. Reliability of low cost, small size and light weighted vehicles and sensors
    • In severe weather conditions, they start to break down and die


Project Achievements

  1. NAF based Architecture
  2. Sensor network (stationary sensors)
    • Reduction of computation time, reduction of communication, and increased robustness to fault organisation
  3. Task allocation (mobile sensing platforms)
    • Developing decentralisation of the developed task allocation algorithm for providing scalability and adaptability
  4. Sensor fusion
    • To propose and incorporate methods that allow the sensors to adaptively assess their own certainty under various operating conditions, and to use this certainty measure to make the fusion output more accurate and robust to unexpected variations in the scene
  5. Information Fusion
    • Reduction of computation time, reduction of communication, and increased robustness to fault organisation
    • Definition of a generic framework for threat assessment providing an analytic representation of behaviour features and a decision policy identifying alleged threats among regular targets
  6. Cooperative FDIR
    • Increased efficiency, allowing reconfigurable network and swarm organisation
  7. Small scale demonstrations
    • Indoor demonstration
    • Outdoor demonstration

Project Information

Partners

  • Cranfield University
  • ONERA, the French Aerospace Lab
  • FOI, Swedish Defence Research Agency
  • University of Patras

Funded by EU-EDA

Info


CONTACT


Vaios Lappas, Research Professor
Department of Mechanical Engineering and Aeronautics
University of Patras
E-mail: vlappas@upatras.gr
Location: Rio, Patras, 26500, Greece