Author name cluster

Shupeng Lai

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers

1 author row

IROS Conference 2023 Conference Paper

An Interactive System for Multiple-Task Linear Temporal Logic Path Planning

Yizhou Chen
Xinyi Wang 0007
Zixuan Guo
Ruoyu Wang 0032
Xunkuai Zhou
Guidong Yang
Shupeng Lai
Ben M. Chen

Beyond programming robots to accomplish a single high-level task at a time, people also hope robots follow instructions and complete a series of tasks while meeting their requirements. This paper presents an interactive software system that consists of a multiple-task linear temporal logic (LTL) path planner and a human-machine interface (HMI). The HMI transforms human oral instructions into task commands that can be understood by the machine. The planner grows a rapid random exploring tree to search for solutions for multiple tasks. When switching tasks, the search tree is re-initialized and reconnected to utilize the information gathered during the exploration of the workspace. The feasibility of the improved planner is theoretically guaranteed, and profiling in simulation shows an acceleration in planning. An experiment with a quadcopter is conducted to show that the combination of the multiple-task LTL planner and the HMI results in a synergistic effect in real-world applications.

Details

IROS Conference 2018 Conference Paper

Optimal Constrained Trajectory Generation for Quadrotors Through Smoothing Splines

Shupeng Lai
Menglu Lan
Ben M. Chen

In this paper, we present a trajectory generation method for quadrotors based on the optimal smoothing B-spline. Compared to existing methods which rely on polynomial splines or time optimal control techniques, our method systematically addresses the issue of axes-coupled and interval-wise constraints. These constraints can be used to construct safe flying zones and satisfy vehicle's physical limits. The proposed approach has also been extended to generate trajectories from the nominal plan which consists of not only points but also lines and planes, opening a door for new improvements and applications. Moreover, a closed-form solution can be obtained for cases without inequality constraints. Such a solution is numerically stable for the large-scale fitting problem, which allows us to directly fit the human sketching input from the touch device and capture all subtle details. Our approach is verified by various real flight experiments. .

Details