论文信息 - On-line MPC-based Stochastic Planning in the Non-Gaussian Belief Space with Non-convex Constraints ∗

On-line MPC-based Stochastic Planning in the Non-Gaussian Belief Space with Non-convex Constraints ∗

A convex program to design a feedback policy in nonGaussian belief space is proposed for a class of systems with differentiable non-linear observation model. A Particle Filter is adopted for belief representation. The size of the optimization vector grows linearly with time horizon and dimension of the state. A new approach is proposed to deal with obstacles as a part of optimization problem. Included results show the efficacy of our method.

Amirhossein Tamjidi | Suman Chakravorty | Mohammadhussein Rafieisakhaei

[1] Nancy M. Amato,et al. FIRM: Sampling-based feedback motion-planning under motion uncertainty and imperfect measurements , 2014, Int. J. Robotics Res..

[2] Robert Platt,et al. Convex Receding Horizon Control in Non-Gaussian Belief Space , 2012, WAFR.

[3] John Lygeros,et al. Stochastic Model Predictive Control using a combination of randomized and robust optimization , 2013, 52nd IEEE Conference on Decision and Control.