March 24, 2014

RoboCupRescue Robot League Research Fields

These are some research fields that we need to have to research about these for our Robocup Rescue - Rescue Robot Team.

1- People Detection
1.1- With Thermal Data
1.2- With 3D Data (from laser or kinect)
1.3- Body part detection
1.4- Vision Based Victim Detection
2D and 3D Human Pose Estimation in Single Images
Pictorial Structures Revisited
People Detection
Pedestrian Detection
People Tracking
Human 3D Pose Estimation
Articulated Pose Estimation
Discriminative Appearance Models for Pictorial Structures
Articulated Pose Estimation and Tracking
Monocular 3D pose estimation and tracking by detection
2D Articulated Human Pose Estimation
Unsupervised Image Classification
Human pose estimation for Kinect
Kinect pose estimation
Detection and recognition of unstructured human activity in unstructured environments
Human body posture detection
Object Detection Technique Development for a Disaster Area

2- The problem of simultaneous localization and mapping (SLAM)
2-1- Google Tango
- Visual SLAM (Visual Simultaneous Localization And Mapping)
- Interest Point Detectors and Local Descriptors for Visual SLAM
- Optical Navigation System
- 3D Simultaneous Mapping and Localization (SLAM) algorithms
- Monocular SLAM / MonoSLAM
- Parallel Tracking and Mapping System
- A Flexible and Scalable SLAM System with Full 3D Motion Estimation

3- Robot Localization
3-1- Inertial Measurement Unit Based Pose Estimation
3-2- RSSI-Based Localization
3-3- Vision-Based Mobile Robot Navigation (Visual Odometry)
3-4- Inertial Sensor Data Integration in Computer Vision Systems
- Robot's Global Positioning
- Robot Pose Estimation - Self Localization
- Scan Matching
- Monte Carlo Localization
- Robot pose estimation in unknown environments by matching 2d range scans
- The normal distributions transform: A new approach to laser scan matching
- A real-time algorithm for mobile robot mapping with applications to multi-robot and 3D mapping
- Monte carlo localization for mobile robots
- Monte carlo localization: Efficient position estimation for mobile robots
- Depth Camera Based Indoor Mobile Robot Localization and Navigation
- Fast Indoor Radio-Map Building for RSSI-based Localization Systems
- Autonomous Mapping and Navigation 
- Robust real-time tracking by fusing measurements from inertial and vision sensors
- vision-based mobile robot navigation using image processing and cell decomposition
- Localization in Wireless Networks via Laser Scanning and Bayesian Compressed Sensing

4- Compare devices for map building, localization, collision avoidance
*LRF (Laser Range Finders)
*LIDAR (Light Detection and Ranging) 
- Are laser scanners replaceable by Kinect sensors in robotic
- Compare 
-- Microsoft Kinect (for Windows / for Xbox 360)
-- Microsoft Kinect for Windows v2  K4W2 
-- Laser Scanner: SICK (like LMS 200)
-- LRF: Hokuyo (like URG-04LX)
-- Asus Xtion / PRO
-- Asus Xtion Live / PRO
-- PrimeSense Carmine

No comments: