Proposal
For
Independent Study project (ISP)
Electronic Engineering Department
Sensor Fusion To Model Dynamic Environment
Hira Amjad
N.E.D University of Engineering & Technology Karachi
Application for ISP
1. Project Identification
A. Project Title:
Sensor Fusion To Model Dynamic Environment
B. Project Supervisor:
Name: Dr. Syed Riaz Un Nabi Jafri
Designation: Assistant Professor
Organization: Department of Electronic Engineering, NEDUET
Mobile # : Tel. # : –
Email: riazun1036@gmail.com
C3. Student Group Lead:
Name: Hira Amjad
Roll No: EL-076/2016-17
Department: Department of Electronic Engineering
Mobile # : 03322363979 Tel. # : –
Email:
hiraamjad06@yahoo.com
2. Scope, Introduction and Background of the Project
A. Scope of the Project:
B. Introduction:
The term SLAM is as stated an acronym for Simultaneous Localization And Mapping.
SLAM is concerned with the problem of building a map of an unknown environment by a mobile robot while at the same time navigating the environment using the map.
SLAM consists of multiple parts; Landmark extraction, data association, state estimation, state update and landmark update.
The SLAM process consists of a number of steps. The goal of the process is to use the environment to update the position of the robot. Since the odometry of the robot (which gives the robots position) is often erroneous we cannot rely directly on the odometry. We can use laser scans of the environment to correct the position of the robot. This is accomplished by extracting features from the environment and reobserving when the robot moves around. An EKF (Extended Kalman Filter) is the heart of the SLAM process. It is responsible for updating where the robot thinks it is based on these features. A unified EKF SLAM method
is introduced and tested in an indoor place which is using 2D laser sensor with webcam integrated on each robot to perceive the environment by means of line features.
The range measurement device used is usually a laser scanner nowadays. They are very precise, efficient and the output does not require much computation to process.
The first step in the SLAM process is to obtain data about the surroundings of the robot. As we have chosen to use a laser scanner we get laser data. The SICK laser scanner we are using can output range measurements from an angle of 100° or 180°. The output from the laser scanner tells the ranges from right to left in terms of meters.
But our work is based on hand held system which is to integrate camera, 2D laser scanner and inertial measurement unit (IMU) for the measurement of the position. It is used in mapping indoor environment.
B1. Project Background and Literature Review:
Project Background:
This project is based on the real fusion of multiple sensors to establish a compact and accurate model of environment. It is planned to use 2D Laser Scanner with a camera to perceive environment and construct a 2D Map model of the surveyed environment. It is further plan to use inertial measurement unit to get pose of the sensor system and to use a probabilistic algorithm to establish Simulation Localization and Mapping Solution. Its application is in surveying and digitalized information modeling of indoor environment. The proposed system has been tested in an indoor environment. The algorithm is developed in Matlab and many functions are used from different works available as open source.
Multi-mobile robotic systems are in use in different applications and offer advantage of fast operation with more accuracy as compare to single robotic system. The success of the multi robotic system is based on how well they solve localization and mapping (i.e. SLAM) issues during the exploration and keep updated to each other. Each robotic unit could consider as a dynamic sensor node which is capable of measuring environmental information. So a moving multimobile robotic team has a good tendency to explore an unknown environment with no fixed sensor network and produces informational structured space. Many research works are based on perceiving 2D information representation of the environment.
From the laser scans, we are extracting(horizontal) lines and corners and from images, we are extracting only vertical lines.laser sensor is range bearing sensor but camera is a bearing only sensor.so features input from both sensors are different in nature.
2D laser scanner are providing raw range values of the surrounding at specified bearings which then convert into geometrical entities using suitable algorithm. They provide both range and bearing as compare to camera which is providing bearing only but with limitation of providing 2D information. We are using laser based horizontal geometrical (line and corner) features which are already used in many state of art work.
Literature Review:
There are many research works have been published in this domain. Therefore in a recent development [1] authors have shown a multi mobile robot simultaneous localization and mapping (SLAM) system for feature based environment learning by using team of exploring robots. Each robot is equipped with 2D laser scanner and a webcam and it serves as a moving sensor node to perceive horizontal and vertical line features respectively.
[2] Data association is critical for Simultaneous Localization and Mapping (SLAM). In a real environment, dynamic obstacles will lead to false data associations which compromise SLAM results. A hybrid approach of data association based on local maps by combining ICNN and JCBB algorithms is used initially.
[3] Development of the information-fusion methods for mobile robots performing simultaneous localization and mapping (SLAM) adapting search and rescue (SAR) environment. Fusion systems consist of laser range finder (LRF) sensors, localization sonars, gyro odometry, Kinect-sensor, RGBD camera, and other proprioceptive sensors.
[4] Proposed a survey of the Simultaneous Localization And Mapping (SLAM) field when considering the recent evolution of autonomous driving. The growing interest regarding self-driving cars has given new directions to localization and mapping techniques.
[5] Presents a multi-sensor fusion based fault detection and isolation (FDI) method for robotic system. The proposed approach is independent of analytical model which suits dynamic, uncertain and unstructured nature of the robot operating environment.
[6] Proposed a novel fusion positioning strategy for land vehicles in GPS-denied environments, which enhances the positioning performance simultaneously from the sensor and methodology levels. It integrates multiple complementary low-cost sensors not only incorporating GPS and microelectromechanical-based inertial measurement unit, but also a “virtual” sensor, i.e., a sliding-mode observer (SMO).
[7] Present a novel algorithm for mobile robot localization and mapping based on stereo vision. This proposed method based on the PLOT features achieves localization and mapping for mobile robot with higher precision.
[8] Presents multi-robot simultaneous localization and mapping (SLAM) framework for a team of robots with unknown initial poses. The proposed solution is using feature based Rao-Blackwellised particle filter (RBPF) SLAM for each robot working in an unknown environment equipped only with 2D range sensor and communication module.
There are many products works like Zebedee, navvis and viametris. But our work is based on Zebedee which is hand held device that consists of a lightweight LiDAR scanner with 30m (100ft) maximum range and a industrial-grade MEMS inertial measurement unit (IMU) mounted on a simple spring mechanism. Zebedee’s ability to self-localize enables its use in indoor, underground, and otherwise covered environments such as dense forest.
Another product is The NavVis M3 Trolley is a fully comprehensive indoor data capture device. Designed for everyday scanning work, the M3 is lightweight and can be disassembled and reassembled within minutes for transportation and convenient storage. It is used in 3D scanning.
Viametris is also a product. It founded by Jérôme Ninot December 2007. Viametris aims to bring Mobile Mapping Solutions to professionals in need for good and reliable tools the innovative way. First a dedicated Research and Development team working with passion, now we design mapping solutions.
B2. Current State of the Art:
In the field of Multi-mobile robotic systems are in use in different applications and offer advantage of fast operation with more accuracy as compare to single robotic system. To represent the environment in compact form, line and corner features (or point features) are used. By sharing and comparing distinct feature based maps of each robot, a global map with known poses is formed without any physical meeting among the robots. This approach can easily applicable to the distributed or centralized robotic systems with ease of data handling and reduced computational cost.
Another approach is Zebedee Product which is hand held device. Zebedee rapidly generates a 3D map as the operator walks through a site. The laser scanner rocks back and forward on a spring, continuously scanning the environment, while a computer records the sensor data. The system offer a unique combination of portability, efficient data collection, accuracy in areas with no GPS, rapid scanning of large areas, and automatic data processing. Zebedee is being used by us and others for mapping mines, heritage sites, and crime scenes.
C. Challenges:
(Please describe the challenges, specific to this research topic, currently being faced internationally.)
D. Motivation and Need:
Multi-mobile robotic systems are in use in different applications and offer advantage of fast operation with more accuracy as compare to single robotic system. The success of the multi robotic system is based on how well they solve localization and mapping (i.e. SLAM) issues during the exploration and keep updated to each other. Each robotic unit could consider as a dynamic sensor node which is capable of measuring environmental information. So a moving multimobile robotic team has a good tendency to explore an unknown environment with no fixed sensor network and produces informational structured space. , that’s why we need such a multiple sensor to establish a compact and accurate model of environment.
E. Project Test Bed – Mention a proposed schema of your test bed with diagrams and subjective definition if necessary
Methodology:
The proposed approach is in indoor environment, geometric shapes are common such as tables, walls which are easy to model. In proposed scheme we are detecting the geometrical shapes from integrated sensors on each robot and extracting lines and features from them to model the environment. Two kinds of integrated sensors are used, 2D laser sensor which senses (horizontally) objects in its view and a camera (a simple webcam) to sense objects in its fronts.
References:-
[1] Syed Riaz un Nabi Jafri, Ryad Chellali, IEEE Member, “A Distributed Multi Robot SLAM System for Environment Learning”
[2] Bai-Fan Chen1, Zi-Xing, Cai2 and Zhi-Rong Zou3, “A Hybrid Data Association Approach for Mobile Robot SLAM”
[3] Hongling Wang and Chengjin Zhang*, Yong Song and Bao Pang, “Information-fusion Based Robot Simultaneous Localization and Mapping Adapted to Search and Rescue Cluttered Environment*”
[4] Guillaume Bresson, Zayed Alsayed, Li Yu, and Sebastien Glaser, “Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving”
[5] A. Abid and M. Tahir Khan,“Multi-sensor, Multi-level Data Fusion and Behavioral analysis based Fault Detection and Isolation in Mobile Robots.”
[6] Xu Li, Member, IEEE, and Qimin Xu, “A Reliable Fusion Positioning Strategy for Land
Vehicles in GPS-Denied Environments Based on Low-Cost Sensors”
[7] Rui Lin, Zhenhua Wang, Rongchuan Sun and Lining Sun, “Vision-based Mobile Robot Localization and Mapping Using the PLOT Features”
[8] Syed Riaz un Nabi Jafri1,3, Zhao Li1 , Aftab Ahmed Chandio2 and Ryad Chellali1 , “Laser only Feature based Multi Robot SLAM”
https://contest.techbriefs.com/2012/entries/machinery-and-equipment/3050
https://data61.csiro.au/en/Our-Work/Monitoring-the-Environment/Sensing-the-environment/Zebedee
http://www.navvis.com/