Skip to content

qinwenchuang/Learning-SQP

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Learning-SQP

==还没研究明白GitHub对LaTeX的支持,ReadMe中公式的显示很乱,可以参考ReadMe.png吧。==

0. 依赖开源库

  • qpOASES
  • Eigen3

1. 编译和运行

2. 初步的简要说明

2.1 LBFGS的计算公式

非线性问题的Hessian矩阵一般很难求解,并且耗时严重,一般会采用Hessian矩阵的近似,考虑到求解耗时和收敛性,优先选择L-BFGS。L-BFGS的计算使用到前后帧的$x$和目标函数的偏导,为了方便描述进行如下简写, $$\begin{align} s_k = x_{k+1}-x_k \end{align}$$ $$\begin{align} y_k = \bigtriangledown_xf(x^{k+1})-\bigtriangledown_xf(x^{k}) \end{align}$$ $$\begin{align} r_k = \theta_ky_k + (1-\theta_k)B_k s_k \end{align}$$ $$\begin{align} \theta_k = \left{\begin{matrix} 1 & if s_k^T y_k \ge 0.2s_k^TB_ks_k\ (0.8s_k^TB_k s_k)/(s_k^TB_k s_k-s_k^Ty_k) & if s_k y_k < 0.2s_k^TB_k s_k \end{matrix}\right. \end{align}$$ $B_k$的更新公式为,$B_k$的更新公式为, $$\begin{align} B_{k+1} = B_k - \frac{B_ks_k s_k^TB_k}{s_k^TB_k s_k} + \frac{r_kr_k^T}{s_k^Tr_k} \end{align}$$

3. TODO

  • 验证积分器的偏导
  • 补全轨迹规划的示例

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published