Motion planning

BACK TO OTHER PROJECTS

WeChat Image_20190626163114

mymap

WeChat Image_20190626163114

1/2

Description

Abstract

•After we have the controller, we need to generate a path planner for it in order to get desired position as an input. What I did is graph search based path planning, which is classic Dijkstra and A star algorithm. This is useful when we have occupancy grid aka voxel map from point cloud. The edge is the cost. The node is the position.

•After we getting the way-points, we should put it into a trajectory generator, adding motion control constrain onto it, which is a process called trajectory smoothing. We use Euler Lagrange Equation to minimize the running cost.

•Minimum snap Trajectory The minimum-snap trajectory is a 7th order polynomial

•We have n points, we have 8n constrains, we linearize it, and put it together to generate a matrix. AX=B, B is the constrain, x is the coefficient.