Open JanMuller1019 opened 4 months ago
您好，去想请问一下，在您开源的代码里，粗跟踪中，IMU的正规方程雅可比为何是这样子的 Mat44 T_temp = T_BC.matrix()T_WD.matrix()M_DCj.inverse(); Mat66 J_rel = (-1Sim3(T_temp).Adj()).block(0,0,6,6); // Mat44 T_temp = T_BC.matrix()M_DCj.inverse(); // Mat66 J_rel = (-1SE3(T_temp).Adj()); Mat66 J_xi_tw_th = SE3(M_DCi).Adj(); // LOG(INFO)<<"-1Sim3(T_temp).Adj: "<<-1*Sim3(T_temp).Adj().matrix();
// LOG(INFO)<<"J_imu1: \n"<<J_imu1;
// Mat99 J_rel = Mat99::Identity(); // J_rel.block(0,0,6,6) = M_rel; Mat66 J_xi_r_l = refToNew.Adj().inverse(); Mat66 J_2 = Mat66::Zero(); J_2 = J_imu1J_relJ_xi_tw_th*J_xi_r_l;
就是最后这个J_2为什么是那四个分量构成的，第一个J_imu为什么只包括了旋转和位置的，能否请您解释一下呢，谢谢您
您好，去想请问一下，在您开源的代码里，粗跟踪中，IMU的正规方程雅可比为何是这样子的 Mat44 T_temp = T_BC.matrix()T_WD.matrix()M_DCj.inverse(); Mat66 J_rel = (-1Sim3(T_temp).Adj()).block(0,0,6,6); // Mat44 T_temp = T_BC.matrix()M_DCj.inverse(); // Mat66 J_rel = (-1SE3(T_temp).Adj()); Mat66 J_xi_tw_th = SE3(M_DCi).Adj(); // LOG(INFO)<<"-1Sim3(T_temp).Adj: "<<-1*Sim3(T_temp).Adj().matrix();
// LOG(INFO)<<"J_imu1: \n"<<J_imu1;
// Mat99 J_rel = Mat99::Identity(); // J_rel.block(0,0,6,6) = M_rel; Mat66 J_xi_r_l = refToNew.Adj().inverse(); Mat66 J_2 = Mat66::Zero(); J_2 = J_imu1J_relJ_xi_tw_th*J_xi_r_l;