有效载荷(计算)
控制理论(社会学)
计算机科学
避障
计算
运动规划
平坦度(宇宙学)
搭配(遥感)
差速器(机械装置)
混合动力系统
控制工程
数学优化
数学
算法
工程类
人工智能
机器人
航空航天工程
移动机器人
物理
机器学习
量子力学
计算机网络
宇宙学
网络数据包
控制(管理)
作者
Jun Zeng,Prasanth Kotaru,Mark W. Mueller,Koushil Sreenath
出处
期刊:IEEE robotics and automation letters
日期:2020-02-11
卷期号:5 (2): 3074-3081
被引量:43
标识
DOI:10.1109/lra.2020.2972845
摘要
Generating agile maneuvers for a quadrotor with a cable-suspended load is a challenging problem. State-of-the-art approaches often need significant computation time and complex parameter tuning. We use a coordinate-free geometric formulation and exploit a differential flatness based hybrid model of a quadrotor with a cable-suspended payload. We perform direct collocation on the differentially-flat hybrid system, and use complementarity constraints to avoid specifying hybrid mode sequences. The non-differentiable obstacle avoidance constraints are reformulated using dual variables, resulting in smooth constraints. We show that our approach has lower computational time than the state-of-the art and guarantees feasibility of the trajectory with respect to both the system dynamics and input constraints without the need to tune lots of parameters. We validate our approach on a variety of tasks in both simulations and experiments, including navigation through waypoints and obstacle avoidance.
科研通智能强力驱动
Strongly Powered by AbleSci AI