第十章作业

基于图优化的滑动窗模型包含：

所以首先需要推导各残差关于优化变量的雅可比

1. 推导各残差关于优化变量的雅可比

Residual

$$ e_p = R_i^T(t_j-t_i) - t_{obs} $$

$$ e_{lie} = ln(R_{ij_obs}^T\cdot(R_i^T\cdot R_j))^{\vee} $$
Jacobian

$$ \begin{aligned} \frac{\partial e_p}{\partial{t_i}} = -R_i^T \end{aligned} $$

$$ \frac{\partial{e_p}}{\partial{t_j}} = R_i^T $$

$$ \begin{aligned} \frac{\partial{e_{lie}}}{\partial{R_i}} = \end{aligned} $$

$$ \begin{aligned} \frac{\partial{e_{lie}}}{\partial{R_j}} = \end{aligned} $$

作业是参考了葛垚大佬的推导，先完成了代码补全，运行起来看了一下效果

	Optimized	LidarOdometry
Map