Having ample practical experience of almost 13 years in the field of robotics makes me a viable candidate for carrying out the research work, if given the opportunity. It would also enable me impart my knowledge to the development of the field of robotics, with practical application to the developing industry in Pakistan.
Atif Anwer
PhD Student, Roboticist, Geek
Research
Select Research Projects
-
This work presents the use of Microsoft Kinect, a commercial RGB-D camera, for a small-scale real-time underwater object 3D scene reconstruction and 3D scanning. Using commercially available depth sensors as a cost-effective alternative, this work successfully shows that real-time underwater 3D scene reconstruction is possible using commercial depth cameras with acceptable trade-offs on accuracy and scanning distance. The acquired depth data is processed to accommodate effects of refraction of the sensor housing and water. The errors in depth calculation due to the change of medium for a time of flight sensor are rectified, in addition to the removal of extra noise faced in underwater imaging, without significant loss of features. This pre-processed data is then passed on to a custom implementation of Kinect Fusion for 3D object scanning and model creation using the Kinect sensor. The experimental findings give promising results that the proposed method can reconstruct 3D surface of underwater objects accurately and in real-time. (Video Summary of the Research)
MSc. THESIS FINAL PRESENTATION:
<iframe src=”https://prezi.com/view/FA2bfYjE4bZGapnCmCJv/embed” width=”550″ height=”400″> </iframe> -
Design of a customized EEG Headset, with design features such as onboard electronics, customized EEG electrodes and variable location measurements. Concepts such as DFAM (Design for Additive Manufacturing) were evaluated and tested for a 3D printed prototype assembly for a client.
-
The project implements real-me object localizaon, SURF feature tracking (using KLT), and PID control implemented on a Lego Mindstorm robot using MATLAB and a monocular camera setup. The robot is successfully follows the person in an unstructured environment avoiding any obstacles in the way.
-
Started in 2014, Scifacterz was a Robotics Startup venture centered on cutting edge projects in various domains. The main project was “NBE (Non Biological Entity); an opensource, low cost, research platform which doubles as a feature packed Home Service Robot”. We were the first round favorites for ”Robot Launch 2014”, the first global startup competition for robotics held by RoboHub.com in collaboration with Silicon Valley Robotics
Key Features include:
- OpenSource ROS package and Android application for easy control and access to all sensors, drive controller, RGB-D / Pointcloud data.
- Easy access to parts/components such as battery pack, drive controller, vision system (stereo cameras & Kinect) alongwith all on-board sensors/electronics to allow painless swapping and hacking.
- Modifiable and updateable drive controller allowing testing of novel control, navigation, obstacle avoidance and collision Detection algorithms.
- Rapidly implement, test and analyse computer vision, machine learning algorithms via integration of Matlab and ROS on any laptop connected to the Robot.
- Space onboard to connect laptop; allowing unrestricted movement while testing algorithms.
- Rugged.Modular.chassis based design with easy assembly/disassembly enabling students to tinker around with hardware.
- Aesthetically designed user friendly high-end version also available.
Details and Images can be found at Scifacterz Website
-
Started in 2014, Scifacterz is a Robotics Startup venture centered on cutting edge projects in various domains, enabling us to turn Science-Fiction to Science-Fact. NBE (Non Biological Entity); is aimed to be an opensource, low cost, research platform which doubles as a feature packed Home Service Robot”.
The Design stereo cameras and multiple sensors to develop a feature packed research platform, with two versions. The NBE MARK-I is a modular learning and experimentation platform geared towards academia from high school to undergraduate/graduate students.
The NBE PRIME is an eye catching design built around the modular chassis of NBE MARK-I & designed to be a feature packed multi-function home service robot.
The NBE was also selected as first round favourite for Robot Launch 2014 . NBE was also presented in Jumpstart Pakistan event in April 2014 and WFAR 7 workshop, held in Lahore in October 2014. -
The project implements SURF feature detecon algorithm in MATLAB using SURFmex (OpenCV implementaon of SURF in Matlab). Monochrome Glyphs were successfully
detected in RGB videos recorded for the project. -
Various feature detecon like SURF, SIFT and pattern recognition techniques were implemented to detect cars (side view only) in colour images.
-
The project was a medium sized 6-DOF robotic manipulator for industrial applications. The manipulator was designed manufactured locally and forward & inverse kinematics based control algorithms were implemented through an ndustrial grade embedded controller (C8051F06x) interfaced with an HMI Device, and AVR μC based servo control.
-
The project was successfully completed at Center of Advanced Studies in Engineering (C@SE) by Undergraduate Students of Electrical Engineering in 2010. The robot was a GPS guided Semi-autonomous all-terrain vehicle, The mechanical design was inspired version of the Shrimp Rover (original design by Institute of Robotics and Intelligent Systems, Switzerland
-
The aim of the project was to build a robot that is capable of mimicking the Human Walking Gait. Emphasis was given to the designing of a stable walking mechanism that can be controlled using standard control algorithms. The mechanical design was novel utilizing Hoeken Straight line mechanism based design for achieving the human gait.