control loop

[zhuyeyiyi]

In STIVS_controller.cpp file, the controller loop (line 83-91) is designed as follows, NUM_STEPS = 5 for(iter = 0; iter < NUM_STEPS; ++iter){ acado_feedbackStep(); if(acado_getKKT() < KKT_THRESHOLD) break; acado_preparationStep(); }

I think this loop is redundant. The measurements keep invariant during this loop. One time is enough { acado_feedbackStep(); acado_preparationStep(); } .

Additional iteration will increase the computational complexity.

[yaofulan]

In STIVS_controller.cpp file, the controller loop (line 83-91) is designed as follows, NUM_STEPS = 5 for(iter = 0; iter < NUM_STEPS; ++iter){ acado_feedbackStep(); if(acado_getKKT() < KKT_THRESHOLD) break; acado_preparationStep(); }

I think this loop is redundant. The measurements keep invariant during this loop. One time is enough { acado_feedbackStep(); acado_preparationStep(); } .

Additional iteration will increase the computational complexity.

你好，我也发现了这个问题，多次循环求解控制量是不变的，这是为什么呀。我的问题是我的求解精度不够，所以想通过循环求解来减少acado_getKKT（acado_getKKT越小越收敛），但是我发现循环过程中acado_getKKT确实能减少，然后控制量始终不变，我很疑惑