Post on 20-Jan-2017
Monocular Simultaneous Localisation and Mapping (SLAM)
Xuechen LiuSupervisor: Simon
MaskellAssessor: Richard
WilliamsDepartment of Electrical Engineering and Electronics
University of LiverpoolMail: sgxliu14@student.liv.ac.uk
What is SLAM: Chicken or Egg
Xuechen Liu sgxliu14@student.liv.ac.uk
A map is asked to be built for localization……
While the pose of the robot need to be estimated based on accurate mapping
Where am I?
What surrounds me?
SLAM Market
Xuechen Liu sgxliu14@student.liv.ac.uk
Autonomous vehicles
Portable Devices
Conventional Approach: KF Based
Xuechen Liu sgxliu14@student.liv.ac.uk
All Gaussian……N !
All Linear!Motion
Measurement
Background: Particle Filter & FastSLAM2.0
Xuechen Liu sgxliu14@student.liv.ac.uk
Generic Particle Filter
Prediction
Weight Assignment
Resampling
Iterative Process Essence of
FastSLAM2.0Through
Particles
Contribution: UKF Replaces EKF
Xuechen Liu sgxliu14@student.liv.ac.uk
Both two Filters can model non-linear distribution, but……EKF(Extended Kalman Filter)
UKF(Unscented Kalman Filter)
V.
S
Taylor Expansion
Unscented Transform
UKF has
BETTER estimation than EKF
UKF: Better Estimation, Easier Understanding, Harder Implementation!
Contribution: Prior Vision System
Xuechen Liu sgxliu14@student.liv.ac.uk
Generally Speaking……Read Video
Corner Detection & Matching
Reconstruct 3D points
Pose EstimationFAST
Corner Detection
Structure from Motion
Final System: How it Will Work?
Xuechen Liu sgxliu14@student.liv.ac.uk
WE WILL SEE.
Further Actions
Xuechen Liu sgxliu14@student.liv.ac.uk
1) Implement the whole system
2) Gaussian Removal: Get the exact expression of weights back
3) Student T likelihood model application [optional]
They can work well separately, so how about the combination?
Literature Review/References[1] M. Montemerlo, S. Thrun, D. Koller and B. Wehbreit, "FastSLAM: A Factored Solution to the
Simultaneous Localization and Mapping Problem," 2002.
[2] M. Montemerlo, S. Thrun, D. Koller and B. Wegbreit, "FastSLAM2.0: An Improved Particle Filtering Algorithm for Simultaneous Localization and Mapping that Provably Converges," in International Joint Conference on Artificial Intelligence, 2003.
[3] M. Arulampalam, S. Maskell, N. Gordon and T. Clapp, “A Tutorial on particle Filters for Online Nonlinear/Non-Gaussian Bayesian Tracking,” IEEE Transactions on Signal Processing, vol. 50, no. 2, pp. 174-187, 2002.
[4] F. Dellaert, D. Fox, W. Burgard and S. Thrun, "Monte Carlo Localization for Mobile Robots," in International Conference on Robotics & Automation, Detroit, Michigan, 1999.
[5] H. Bay, A. Ess, T. Tuytelaars and L. Van Gool, “Speed-Up Robust Features (SURF),” Computer Vision and Image Understanding, no. 110, pp. 346-359, 2007.
[6] E. Rosten, R. Porter and T. Drummond, "Faster and Better: A Machine Learning Approach to Corner Detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 32, no. 1, pp. 105-119, 2010.
[7] S. Julier and J. Uhlmann, “Unscented Filtering and Nonlinear Estimation,” Proceedings of the IEEE, vol. 92, no. 3, pp. 401-422, 2004.
Xuechen Liu sgxliu14@student.liv.ac.uk
Thanks for Listening!Any Questions?
Xuechen Liu sgxliu14@student.liv.ac.uk