Sensor systems for urban operations - Annual report 2010

Authors:

  • Stefan Nilsson
  • Staffan Abrahamson
  • Maria Andersson
  • Thord Andersson
  • Erika Bilock
  • Hans Habberstad
  • Fredrik Hemström
  • Fredrik Lantz
  • Dietmar Letalick
  • Staffan Lindström
  • Henrik Petersson
  • Joakim Rydell
  • Karl-Göran Stenborg

Publish date: 2010-12-27

Report number: FOI-R--3096--SE

Pages: 25

Written in: Swedish

Keywords:

  • sensor systems
  • sensor networks
  • multi-sensor fusion
  • urban scenario
  • anomalies
  • anomalies detection
  • continuous surveillance
  • blue force tracking
  • crowd control

Abstract

This annual report gives an account of the activities carried out and the results produced in the three-year Armed Forces' project Sensor Systems for Urban Operations. The aim of the project is to study and evaluate how various combinations of sensor systems can be made to co-operate in order to contribute to an improved situation awareness in the urban environment. The research efforts in the project are focused towards the following important military need areas: Continuous Surveillance, Crowd Surveillance and Blue Force Tracking. Multisensor data fusion, sensors in co-operation, detection/tracking/classification of human movement, and anomalies detection are examples of research areas where efforts are made in the project. The project participates in international collaboration. Through co-financing of the three EU FP7 projects Prometheus, ADABTS, and IMSK, the project gets access to the research and the results generated in these. Furthermore, the project finances co-operation in the NATO group SET-153 RTG-083 Multi-sensor integration in urban operations. In the autumn of 2011 a common multi-sensor experiment will be made in Bonnland, Germany. In Continuous Surveillance we have this year improved our multi-target tracker for tracking of humans. A volumetric method which uses co-operating cameras/views in order to create a volume conception of the scene has been developed. In this way, one can infer what objects conceal what other objects. The method makes it possible to present target tracks from each camera in a common image, together with the associated uncertainty of the estimated target position. Work on the further development of the attribute detector is going on, and together with the volumetric method it is expected to give a more robust track following of temporarily concealed persons, or when people move between different cameras. In a partial study an investigation is made of how good positioning of acoustic objects is achievable with microphone arrays in an urban environment. The sub-project Crowd Surveillance studies and develops methods for detection and tracking of the crowd as a whole on the one hand and single individuals being part of it on the other. A method has been developed which estimates the probability of normal behaviour of a crowd, and that detects deviation from the expected normal state. Data from simultaneous recordings with combinations of visual cameras, IR cameras and acoustic arrays have formed the basis of this development. A special evaluation of 3D laser as being one part of a multisensor system shows that the range information obtained enables improved detection and tracking of humans and an increased robustness in poor light conditions. In the sub-project Anomalies detection in urban environment work has proceeded with the study and development of methods and system concepts for automatic detection of interesting, deviating events and courses of events in urban environment. A framework for fusion of detections of anomalies with a view to detecting interesting events has been developed. Among other things, the framework enables simultaneous handling of various forms of attributes and target tracks, and enables fusion of detections over time. The sub-project Blue force tracking investigates possible techniques for the identification and tracking of the movements of one's own troop in the urban scenario, especially in buildings without GPS coverage. We have experimentally pointed to the possibility of using soldiercarried optical systems for determining the soldier's own position, and simultaneously to construct a map of the interior of the building through the use of the SLAM method. A completely new "silent" method for positioning of one's own soldiers in buildings is being studied, based on a combination of RF/RFID. A test system has been developed here and the method will now be validated experimentally.