Multiple Pursuer-Based Intercept via Forward Stochastic Reachability

We discuss the multiple pursuer-based intercept of a threat unmanned aerial system (UAS) with stochastic dynamics via multiple pursuing UASs, using forward stochastic reachability and receding horizon control techniques. We formulate a stochastic model for the threat that can emulate the potentially...

Full description

Saved in:
Bibliographic Details
Main Authors: Vinod, Abraham P., HomChaudhuri, Baisravan, Hintz, Christoph, Parikh, Anup, Buerger, Stephen P., Oishi, Meeko M. K., Brunson, Greg, Ahmad, Shakeeb, Fierro, Rafael
Format: Conference Proceeding
Language:English
Subjects:
Computational modeling
Convex functions
Fourier transforms
Optimization
Real-time systems
Sensors
Stochastic processes
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We discuss the multiple pursuer-based intercept of a threat unmanned aerial system (UAS) with stochastic dynamics via multiple pursuing UASs, using forward stochastic reachability and receding horizon control techniques. We formulate a stochastic model for the threat that can emulate the potentially adversarial behavior and is amenable to the existing scalable results in forward stochastic reachability literature. The optimal state for the intercept for each individual pursuer is obtained via a log-concave optimization problem, and the open-loop control paths are obtained via a convex optimization problem. With stochasticity modeled as a Gaussian process, we can approximate the optimization problem as a quadratic program, to enable real-time path planning. We also incorporate real-time sensing into the path planning by using a receding horizon controller, to improve the intercept probabilities. We validate the proposed framework via hardware experiments.
ISSN:2378-5861
DOI:10.23919/ACC.2018.8431308