What steps will reproduce the problem?
1. If LinearKalman.refreshProblem() and LinearKalman.step() are not called
exactly once per new data node, then the linear least squares operation inside
of step() will either skip data or run multiple times on each node, resulting
in an incorrect estimate.
To fix this, Default.LinearKalman should process on-diagonal edges from each
measure using a FIFO to store indices, thereby keeping track of incoming data.
Original issue reported on code.google.com by dddvis...@gmail.com on 20 Sep 2010 at 10:53
Original issue reported on code.google.com by
dddvis...@gmail.com
on 20 Sep 2010 at 10:53