Sensor systems for urban environment - Final report

Authors:

  • Stefan Nilsson
  • Maria Andersson
  • Thord Andersson
  • Fredrik Lantz
  • Dietmar Letalick
  • David Lindgren

Publish date: 2009-01-08

Report number: FOI-R--2614--SE

Pages: 79

Written in: Swedish

Keywords:

  • Sensor Systems
  • Sensors in Network
  • Urban Environment
  • Anomalies Detection
  • Situation Picture
  • Continuous Surveillance
  • Red and Blue Force Tracking
  • Crowd Control

Abstract

This final report gives an account of the activities carried out and the result produced in the three-year Armed Forces' project Sensor Systems for Urban Environment. The aim of the project has been to study and evaluate how various combinations of sensor systems can contribute to an improved military situation awareness in the urban environment, compared to the achievement of the single sensors. In co-operation with military experts four important military need areas were identified, viz. Continuous Surveillance, Crowd Control, Red and Blue Force Tracking, and Gathering of Situation Picture, demanding advanced technical support. Essentially, the project has concentrated its activities towards these areas. The capacity to detect anomalies from the normal picture is a central one, which also has been devoted special research effort. The project is carried out in international co-operation, for instance with projects within the security research program of the EU. Within the sub-projects Continuous Surveillance, Crowd Control and Anomalies Detection measuring campaigns with various multi-sensor systems have been carried out. The aim of the field trials has been to use the collected data for technique evaluation and demonstration in scenarios characterized by violence or potential violence. Measured data has been used to demonstrate how individuals can be detected automatically, be associated and followed in a camera network with limits on communication and distributed algorithms. Acoustical sensors with high spatial resolution have been used to filter out sound sources, enabling sounds to be automatically connected with people or vehicles, monitored with some high-resolution optical sensor. Based on our measured data, a fusion model is proposed for co-located IR and video, enabling more robust detection and tracking of people in future surveillance systems. The fusion model utilizes the richness in colour and detail of the video in combination with the heat radiation sensitivity of the IR-camera, but its insensitivity to reflections, shadows and darkness. The sub-project Anomalies Detection in Urban Environment has been working with the study and development of methods and system concepts for the automatic finding of interesting, anomalous events and courses of events in an urban environment. At LedDoV 2008 in Enköping, the project demonstrated the functionality of a temporarily assembled sensor network for urban surveillance, consisting of hundreds of sensors (visual, IR, acoustic, radar). The sensors work in co-operation locally, and specially developed fusion algorithms continuously generate target tracks. The operator is alarmed when events occur, departing from the normal picture. The communication in the network is effected via a narrow-band radio link where the information is transferred through multi-hop. The abilities Continuous Surveillance, Anomalies Detection and Weapons Detection were demonstrated conceptually. The intent here was to illustrate how an operator uses the sensor system to be aware of potential danger, and to analyse the events behind the alarms. A specially developed operator interface gives the operator the possibility to order target tracks, video or visual images, and to search for objects, which can be used for forensic analysis. In the sub-project Red and Blue Force Tracking an overview is given of possible techniques to identify and track the movement of one's own troop in urban environment and especially in buildings without GPS coverage. The possibility for a soldier to utilize his/her own optical sensors to determine his/her own position, and simultaneously to construct a map of the interior of the building through using the SLAM method, has been studied and judged to be very interesting. Furthermore, a proposal has been made of an entirely new positioning method of one's own soldiers in buildings, based on semi-active RFID.