Project 8: Local Positioning Systems

Leader: K. Åström, LU

Participants at LiU: F. Gustafsson
Participants at LU: B. Bernhardsson, F. Tufvesson

Project description: Navigation, localization and map making are interesting research areas with many interesting applications. Indoor navigation is much more challenging compared to outdoor navigation since there is usually no GPS coverage, electronic compasses do not work that well, and the radio environment is complicated. Visual localization with respect to landmarks has the potential to overcome many of these problems, but needs considerable development to be robust and to be able to scale to large image collections. In this sub-project we will work on several aspects of navigation, localization and map making and investigate the use of combining several different sensor modalities such as cameras, inertial sensors (accelerometers and gyroscopes), Ultra-wideband (UWB), sound, wifi signal strength, wifi round-trip time measurements, field strength measurements).

During 2016, we will among other things work further on new roundtrip time measurements that will come as a
standard in future wifi. We have access to such equipment and drivers so that we can experiment with this new
technology. We will also integrate techniques developed during the first years of ELLIIT, so that they can be tested for different use cases such as human indoor navigation and for autonomous navigation.

  • Joint positioning and radio channel estimation/prediction. We see interesting potentials of using advanced
    positioning techniques for improving radio channel prediction for MIMO systems. The work ranges from
    theoretical analysis of fundamental limits for estimation performance using noisy IMU sensors and radio
    receivers as well as to design of practical fusion algorithms and experimental evaluation. Estimation of timeof-
     arrival and received signal strength measurements and covariance estimates from radio channel estimates.

  • Cellular phase based positioning. In cellular systems the bandwidth is usually confined to 20-40 MHz. In
    order to get accurate positioning, well below 1 m, in such systems it is necessary to utilize all the information
    carried by the radio signal, including its phase, information from the reflections, and possibly information
    about the environment. We will study phase based approaches for highly accurate cellular positioning, with
    the aim of achieving cm level positioning accuracy using only moderate bandwidths.

  • Joint estimation of room geometry, microphone positioning and sound source positioning. We have previously
    developed efficient solvers for several so called minimal problems. There is now potential in developing robust
    algorithms for joint estimation of room geometry, microphone calibration and sound source positioning. Work
    here ranges from theoretical work on minimal problems with missing data to system development and testing.
    This could be used to develop cheap and simple indoor positioning systems that could be used for autonomous
    vehicle positioning and navigation.