commit 39f26f66eedcdec248c7bebe953350aba334a0d7
author Austin Schuh <austin.linux@gmail.com> Thu Feb 24 21:34:46 2022 -0800
committer Austin Schuh <austin.linux@gmail.com> Sat Feb 26 15:51:49 2022 -0800
tree 87f1e7722fa084633bce5eb6d9a8a1250bcd6ecd
parent e2f40d7d59c43b2c827a39569930810af66e8850

Add catapult MPC and controller class

We want to shoot with the MPC, but we want to decelerate and reset with
a more traditional controller.  So, transition to the MPC, and back to a
profiled subsystem.

This makes some tweaks to the solver to get it to converge more
reliably.  There's apparently a scale factor which was scaling down the
cost matrices based on the initial p matrix, causing it to not solve
reliably...  Good fun.

Change-Id: I721eeaf0b214f8f03cad3112acbef1477671e533
Signed-off-by: Austin Schuh <austin.linux@gmail.com>

y2022/control_loops/superstructure/catapult/BUILD

diff --git a/y2022/control_loops/superstructure/catapult/BUILD b/y2022/control_loops/superstructure/catapult/BUILD
index 289dfde..a43925e 100644
--- a/y2022/control_loops/superstructure/catapult/BUILD
+++ b/y2022/control_loops/superstructure/catapult/BUILD
@@ -42,6 +42,10 @@
         ":catapult_plants",
         "//aos:realtime",
         "//third_party/osqp-cpp",
+        "//y2022:constants",
+        "//y2022/control_loops/superstructure:superstructure_goal_fbs",
+        "//y2022/control_loops/superstructure:superstructure_position_fbs",
+        "//y2022/control_loops/superstructure:superstructure_status_fbs",
     ],
 )