can recommend some ways or ideals. we want to modify the trajectory in following figure. and assure trajectory integrate

Question

description: We have an outdoor mobile robot, use RTK+imu+odom algorithm to location and mapping. when RTK is OK, the effect is ok. when no RTK, we use imu+odom to location.

problem: As shown in the following figure, while no RTK, we use odom and imu to calculate trajectory, but the trajectory deviates from the actual trajectory. We hope find a way to modify the trajectory, when RTK recovery. and assure the trajectory is continuous, no break. Please recommend some ways to us.

available data:

1. when have RTK, the real trajectory position(x,y)
2. odom+imu calculate trajectory(x, y), when RTK recover, we still can get odom+imu trajectory

diagrammatic sketch

caption:

1. black line is robot trajectory calculate by our mapping algorithm
2. when have RTK, robot trajectory is the same as real trajectory

we want to modify the trajectory in figure. and assure trajectory integrate.

try: we try VSlam sim3 algorithm, but we find it have some difference. and we don't assure if sim3 algorithm can solve our problem. so we want to find if have other ways to slove it

Answer 1

In your picture, you can assume that when you don't have the RTK, the robot should maintain its rotation till it gets RTK back. This will works well when there a linear reference trajectory.