An improved algorithm based on rapidly-exploring random tree（ RRT） connect to solve the trajectory planning problem of quadrotor unmanned aerial vehicle（ UAV）

which autonomous cruising in low altitude complex environment is proposed. According to differential flat characteristics of quadrotor UAV

design reasonably planning space; introduce angle constraint and path pruning method

on the basis of premising instantaneity of plan

average length of the planning path is shortened by 19 % in the condition of the limited planning time; and through Bspline function for path smoothing

processing to make the generated trajectory has two-order geometric continuity

and it can avoid the collision with obstacle area. Contrast experiments show the validity

superiority and stability of the algorithm.