作者: Pravin Mali , Arun Kumar Singh , Madhav Krishnal , P.B. Sujit
DOI: 10.1109/CASE48305.2020.9216801
关键词:
摘要: In this paper, we consider the problem of tracking a ground vehicle with fixed-wing aerial (FWV) equipped downward-facing camera. The complexity stems from highly nonlinear kinematics FWV and stall speed constraint. We propose Model Predictive Control (MPC) approach for that has two main contributions. Firstly, model requirement through novel constraint function relates FWV’s position orientation to field view Secondly, make case reformulating underlying optimization MPC as an unconstrained solving it state art gradient descent variants like ADAM RMSProp. Specifically, show real-time performance optimizer while achieving good under various kinematic constraints. validate our extensive simulations, specifically highlighting 3D spiral-like trajectories obtained when slow-moving vehicle. also present quantitative analysis efficacy different variants.