Fred159/My-Udacity-Project4-Advanced-lane-finding

Stars
3
Rank 3,945,249 (Top 78 %)
Language
HTML
Created over 6 years ago
Updated almost 5 years ago

Fred159/My-Udacity-Project4-Advanced-lane-finding

Fred159

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

RoboticsND-Follow-me-deeplearning

Jupyter Notebook

3D-Perception

Udacity_robotics_term2_p2_where-Am-I

Monte-Carlo based localization is a widely use algorithm in robot application. The main application in robotics is called particle filter. Particle filter is a kind of filter algorithm which is totally different with filter like KF, EKF, UKF family. Particle filer doesn’t use moment of noise like mean value and variance value to filtering the noise. Particle filter use many particles to approximate many kinds of noise distributions. So particle filter can extract any noise distribution and filtering any noise distribution with enough particles. But as the quantity of particles grows, the more computation power it uses. So it is a trade off of computation efficiency and precision of estimation. In this paper, a ROS package which is named ’AMCL’ with adaptive particle filter is used to localize

SHENLAN_sensorfusion_04

SEHNLANXUEYUAN sensor fusion course homework-chapter 3

My-Udacity-Project1-Lane_Extraction

Udacity project1 -Finding Lane Lines on the Road

Jupyter Notebook

CarND-Term3-Path-Planning-Project

Udacity Self Driving Car path planning project.

My-Udacity-Project3-Clone-Driving-Behavior

Udacity selfdriving car project 3-behavior cloning using deep learning.

My-Udacity-Project5-Vehicle-detection-and-tracking

Udacity_robotics_term2_p3_map_my_world

SLAM is a key problem in robotics. Because of the infinite states and variables, SLAM is a difficult problem. What is more, environment can be dynamic. Once the environment turn in to dynamic, time is also should be considered into algorithm. GraphSLAM is a one of the powerful solutions. It makes a sparse matrix for mapping features and states. The sparse matrix is used for construct constraints for each connections. The reason why use sparse is that system can’t save and do not have to save all the connections relationship. It just keep the important weight about the connections which is directly connected. Actually, the connections are formulated into Gaussian distribution. GraphSLAM has the capability for handling full SLAM problem with the constraints sparse matrix’s help. The simulation system is constructed based on gazebo and rviz.

CarND-Term3-System-Integration-Project

This is the udacity self driving dar nanodegree program Term3. The project is aim to run self driving car based on ROS system. All the communication is based on ROS.

RoboND-Kinematics-Project

Robotics-term2-P4-Arm-Reinforcement-Learning

Robotics term2 project4 arm reinforcement learning project

SHENLAN_sensorfusion_assignment6

SHENLAN_sensorfusion_assignment3

Robotics-P4-pid_controller

SHENLAN_sensorfusion_assignment5

Resources

Some resources from internet that is opened.