In this presentation we will describe novel Bayesian data fusion algorithms for localising cyclists on the velodrome race track. The path of the cyclist is described using a dynamical model expressed in the intrinsic coordinates of the bike, and we show how to fuse this with timing measurements derived from a video capture system and intertial measurements from a gyrometer and accelerometer attached to the bike. A combination of Rao-Blackwellised particle filtering/smoothing and optimal proposal particle filtering is developed for solution of this task.

This presentation is part of Minisymposium “MS74 - Sequential Monte Carlo methods for inverse estimation in imaging science”
organized by: Narayan Puthanmadam Subramaniyam (Aalto University) , Sara Sommariva (Aalto University) .

Authors:

Simon Godsill (University of Cambridge)

Jiaming Liang (University of Cambridge)

Keywords:

bayesian, bayesian methods, imu, intrinsic coordinates, inverse problems, nonlinear optimization, optimal proposal, sequential monte carlo, statistical inverse estimation methods, stochastic processes