Multi-sensor systems for surveillance - Progress report 2015

Authors:

  • Stefan Nilsson
  • Maria Andersson
  • Thord Andersson
  • Erika Bilock
  • Viktor Deleskog
  • David Lindgren
  • Jonas Nygårds
  • Fredrik Näsström
  • Joakim Rydell

Publish date: 2016-04-01

Report number: FOI-R--4240--SE

Pages: 25

Written in: Swedish

Keywords:

  • multi-sensor systems
  • sensor networks
  • detection
  • distributed target tracking
  • classification
  • multi-sensor fusion
  • urban environment
  • urban operations
  • anomaly detection
  • deviation detection
  • surveillance
  • UAV
  • mapping
  • Blue Force Tracking
  • SLAM

Abstract

This progress report gives an account of the activities carried out, and the results produced in the first year of the three-year Swedish Armed Forces sponsored project Multi-sensor surveillance systems. The project has studied how different sensor systems can work together to provide enhanced military situation awareness in the urban environment. The project examines and evaluates how different configurations of collaborative and networkbased multi-sensor systems should be designed to automatically detect, classify and track threats in difficult monitoring situations. The project is also investigating how the data from the individual sensor systems can be distributed to a central unit which creates a common situational awareness, and how multi-sensor systems can be developed for blue force tracking in complex environments. We have also studied how a multi-sensor network should be designed for a robust and sustainable manner to support surveillance in mainly military scenarios. The goal for the year was to implement a study on the topic of distributed sensors and fusion of surveillance applications. Preliminary results exist, but a comprehensive study will be reported next year. The target tracking activities this year have built on the framework that is based on Bayesian statistics and classical target tracking theory. The goal is to complement the existing centralized tracking algorithms for distributed multi-sensor target tracing. A study of existing methods of association and tracking of distributed systems has been carried out in cooperation with activities in the EU FP7 project P5. A particularly interesting method has been selected for further studies. In addition, different algorithms for detection of people in visual images, have been evaluated in a separate study. Within the work package Anomaly Detection, we study how systems for monitoring the urban environment should be designed for early detection of threatening events, particularly in crowds. Methods for anomaly detection have been further developed and evaluated. We have particularly studied how the anomaly detection can be improved by zoning, where the zones, corresponding to various parts of the surveillance area, are characterized by a specific pattern of movement. In simple cases, the activities normally found in the various zones are known from the beginning. In cases where such knowledge is missing, we have shown that zoning can be created automatically by using sensors to build up statistics on the movement patterns. Results of recorded data show that anomalies can be detected at an earlier stage when we take account of zoning. Within the work package Blue Force Tracking, a new real-time positioning and mapping system of unknown indoor environments has been developed. The new system utilizes an active distance measuring sensor, which gives darkness capacity, and significantly improves the system's mapping capability compared to previous solutions with passive stereo cameras. The new system has been evaluated in field trials, for example together with the Swedish Armed Forces Protection Center in Umeå and with the Södertörn Fire Department. The measurements were validated using a reference system and have proven to give good accuracy. Continued cooperation with the Protection Center is planned. In a separate pilot study, we have investigated trends in sensors, sensor platforms and management software that can enable a drone platform to perform independent mapping and monitoring in complex environments. In addition, data collection from two different sensor platforms has been carried out. This data will be the basis for future methodology development, e.g., collision avoidance and mapping. International cooperation is carried out by participation in the two NATO groups Multi-Sensor Integration for Urban Operations and Battlefield Acoustic Sensing, Multi-modal Sensing and Networked Sensing for ISR Applications, and participation in the EU FP7 project P5. This interaction contributes with new knowledge about sensor system configurations for improved situation awareness.