Hongchen Gao

HJIMI Technology Co., Ltd. · R&D Center

The arXiv endorsement E-Mail I received is as below:

(Hongchen Gao should forward this email to someone who's registered as an endorser for the cs.RO (Robotics) subject class of arXiv.)

Hongchen Gao requests your endorsement to submit an article to the cs.RO section of arXiv. To tell us that you would (or would not) like to endorse this person, please visit the following URL:

https://arxiv.org/auth/endorse?x=9GAVSR

If that URL does not work for you, please visit

http://arxiv.org/auth/endorse.php

and enter the following six-digit alphanumeric string:

Endorsement Code: 9GAVSR

Usually, we analyze the observability of a system through an observability matrix. For example, for a VO/VIO system, we can calculate its observability matrix according to its state space equation and analyze its unobservable dimensions: monocular VO is 7, monocular VIO is 4; however, for the monocular VO based on the optimization method, we can also get its unobservable dimension of 7 by calculating the zero space of its Hessian matrix. Is there any connection between them?

I drew a diagram of the SLAM Abstract Model which abstracted and simplied the SLAM procedure, friends who are interested in this are welcome to discuss.

How to add Relocalization function to the open source SLAM framework based on MSCKF (for example: MSCKF-VIO and OpenVINS)? I can get the pose of the relocalization frame through Loop Detection and nonlinear least squares optimization, but the entire system diverges when it is fed back to the MSCKF sliding window.

I read the article Factor Graphs and GTSAM recently in which it defines a unary factor corresponding to a measurement likelihood with a Gaussian noise model as the attached formula.

But, I cannot understand how it is defined.