Multimodal Localization Quad Chart

1
Multimodal Localization Quad Chart

• Lewis Girod (girod_at_lecs.cs.ucla.edu)
• Laboratory for Embedded Collaborative Systems

2
Ad-hoc Deployable Fine-Grained Localization

• Goals
• Minimize deployment requirements
• Localize on a scale with node density
• Environmental independence

3
Components and Challenges

• Components
• Ranging mechanisms (acoustic, optical)
• Synchronization for measuring time of flight
• Multilateration to compute coordinate systems
• Challenges
• Mechanisms to combat drift and cumulative error
• Filtering to discard outliers and data resulting
  from non-line of sight measurements
• Proposed solution Combination of sensor
  modalities

4

• Drift and cumulative error
• Correct using absolute measurements
• Absolute orientation sensor combined with crude
  angle-of-arrival
• Absolute position sensor (GPS or surveyed points)

• Obstructed environments
• Outlier rejection based on lower SNR and
  multimodal distributions
• Geometric consistency checks using AOA to
  sanity-check ranging and find consistent sets of
  nodes
• Cross-validation from different modalities (e.g.
  camera detection gt LOS)