Saumya Kumaar Saksena

Drones Fanatic . Innoventor . Electronics Engineer

About Me

Hey, everyone! Thanks a lot for viewing my profile. I graduated from BITS Pilani, Hyderabad Campus, with Majors in Electronics and Communication from Dept. of Electrical and Electronics Engineering. I am currently working as a Research Assistant at the Department of Aerospace Engineering, Indian Institute of Science. I work on unmanned aerial vehicles (UAVs) and my aim is to make these awesome flying machines more useful by making them autonomous and more reliable. My work revolves around computer vision, deep learning, system identification and brain-computer interface, where I attempt to develop computationally inexpensive algorithms for real time performance.

Keep your eyes on the stars...and your feet on the ground

Research Experience

Indian Institute of Science

Aug'16 - Present

Research Assistant

Autonomous Navigation of UAVs and BCI-based Control of Multi-Rotors

  • Designed and deployed a novel and robust real-time road navigation system for Unmanned Aerial Vehicles. Our algorithm takes 20 ms to process a frame and detect roads, which allows us to process 50 frames per second on ARM architecture. Moreover, the code uses only 16 MB of RAM and 540 MHz of CPU for computation. This makes it easy to deploy in less resourceful computational environments.

  • Developed an algorithm for collision avoidance in cluttered environments based an algo- rithm proposed earlier. Their approach estimates distances to nearest obstacle in various longitudinal stripes of the input image by using a weighted combination of texture features. We have closely studied their algorithm from a theoretical perspective and have identified various major and minor means to enhance the algorithm’s performance. We have validated the modifications incorporated into the original algorithm in a simulated environment as well as in actual autonomous flight experiments.

  • Designed a control methodology for maneuvering a delta-winged UAV using Common Spatial Paradigm (CSP) and Linear Discriminant Analysis (LDA) of the EEG signals and its implementation. These methods, with data processing rate of around 97 Hertz, provide a rapid classification platform for establishing a stable control link to the UAV and also achieve a high level of classification accuracy of 85%.

  • Deployed the prototype showing the control of the human brain interfaced directly with the drone

Flytbase Inc.

May'16 - Aug'16

Research Intern

Precise Collision Avoidance of MAV's Using SONAR (Sound Navigation and Ranging)

  • Designed and integrated a generic framework consisting of 4 (upgradable to 6) ultrasonic sensors to work with FlytPOD, a flight computer. The deisgn is backed by stereovision, which makes collision avoidance easier. The design is made from carbon fibre which provides a sturdier structure to the overall system. The sensor and Arduino holders are 3D printed.

  • Used PS4Eye camera, HC-SR04 ultrasonic sensors and Arduino/TIVA C Launchpad for real time stereovision based collision avoidance. The algorithm so developed is under improvement currently, as only 5 FPS was achieved.

  • Acheived stable flight and precise navigation in low density indoor and outdoor environments.

  • Designed a user-friendly GUI to allow control on Data Rate, Threshold and each Individual Sensor.

It does not matter what I can do.What matters is what I have done.

Industrial Experience

H.B.L. Power Systems, Indian Railways

May'15 - Jul'15

Summer Intern

LVPEI-MIT Media Lab Tech Health Camp

Jul'15 - Aug'15

Project Assistant

  • Designed a cheap, effective and easy to handle automatic ophthalmoscopic device that can detect diseases like cataract, myopia and hypermeteropia. Cost cutting done by an amount of 16K INR.

  • The eye imaging system is based on the red eye e ect, depending on the intensities of the red colour obtained in the image, useful information about the eye could be deduced.

  • Implemented at LV Prasad Eye Institute, this project was implemented at RedX 2015 in association with MIT Media Lab. The prototype has been tested on 25 patients till now.

CETPA Robotics Pvt. Ltd.

May'14 - Jul'14

Summer Intern

  • Designed a prototype for the display board used in metro trains, displays basic information like speed, next station, etc

  • The design was implemented using 8051 uC, LCD board, AVR programmer, done in Keil uVision 4 environment.

  • Project was accepted for real time implementation at Lucknow Metro Rail Corporation, LMRC. Received 1L INR for the same.

Colvin Taluqdar's College

Jul'14 - Aug'14

Teaching Assistant

  • I was a teaching assistant for high school students. Prime tasks were problem setting and taking tutorial classes.

  • I taught fundamentals of Intermediate Physics and Mathematics with special focus on Differential Calculus, Fucntions in Mathematics and Kinematics and Electromagnetism in Physics.

  • I provided assistance to the students in clearing their competitive examinations like IIT-JEE and BITSAT, in which I had already excelled. I also recieved an appreciation certificate for my services.

Alone we can do so littleTogether we can do so much

Research Papers

  • High Speed Autonomous Navigation of Unmanned Aerial Vehicles using novel Road Identification, Following and Tracking (RIFT) Algorithm

    Submiting To : IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2018

    Status : Under Progress

    Autonomous road navigation in unmanned aerial vehicles flying at very low altitudes paves the way for a multi- tude of applications such as security surveillance, monitoring of traffic or pollution, package delivery, ground-vehicle tracking etc. Most research in the field of autonomous road vehicles is focused on lane tracking and other marker-dependent tech- niques. Un-aided junction detection... To read more kindly click on the icon above

    A Performance Study of 14-Channel and 5-Channel EEG Systems for Real-Time Control of Unmanned Aerial Vehicles (UAVs)

    Submitted At : IEEE International Conference on Robotic Computing (IRC) 2018

    Status : Accepted

    Brain-computer interface (BCI), an actively re- searched multi-disciplinary domain, has completely trans- formed the approach to robotic control problems. Researchers have focused on developing algorithms that optimize robotic movement to achieve desired trajectories, and it’s a general understanding that route optimization problems are difficult to solve mathematically. Humans, on the other hand, tend to optimize... To read more kindly click on the icon above

    State-Space Identification of Unmanned Helicopter Dynamics using Invasive Weed Optimization Algorithm on Flight Data

    Submitted At : 6th Asian Australian Rotorcraft Forum (ARF) and Heli-Japan

    Status : Presented

    In order to achieve a good level of autonomy in unmanned helicopters, an accurate replication of vehicle dynamics is required, which is achievable through precise mathematical modeling. This paper aims to identify a parametric state-space system for an unmanned helicopter to a good level of accuracy using Invasive Weed Optimization (IWO) algorithm. The flight data of Align TREX 550 flybarless helicopter is used in the... To read more kindly click on the icon above

  • Vision-based Control for Aerial Obstacle Avoidance in Forest Environments

    Submitted At : IFAC’s International Conference on Advances in Control and Optimization of Dynamical Systems (ACODS) 2018

    Status : Accepted

    The increasing application of unmanned aerial vehicles (UAVs) in unstructured, natural environments raises the demand for robust obstacle avoidance systems that work in real- time. In view of this problem, we present a control algorithm for quadrotors based on monocular vision that is specifically designed for obstacle avoidance in forest environments. The algorithm... To read more kindly click on the icon above

    Brain-computer Interface based Real-time Control of Fixed Wing and Multi-rotor UAVs

    Submitted At : ICACSD 2018 : 20th International Conference on Automatic Control Systems Design

    Status : Accepted

    Brain-computer interfacing (BCI) is a technology that is almost four decades old and it was developed solely for the purpose of developing and enhancing the impact of neuroprosthetics. However, in the recent times, with the commercialization of non-invasive electroencephalogram (EEG) headsets, the technology has seen a wide variety of applications like home automation, wheelchair control, vehicle steering etc. One of the latest... To read more kindly click on the icon above

    DEEP LEARNING APPROACH FOR REAL TIME SPEECH-TO-TEXT-BASED CONTROL OF UAV’s

    Submitted At : International Conference on Information Society and Smart Cities (ISC 2018)

    Status : Accepted

    Speech recognition has been one of the key research domains in computational signal processing. Despite high levels of computational complexity associated with achieving speech recognition in real-time, promising progress has been made under the umbrella of voice controlled robotics. Most speech recognition applications in robotics rely heavily on either... To read more kindly click on the icon above

Either do what you love..or love what you do

Projects

  • Simultaneous Localization and Mapping Using SF-11 LIDAR

    SLAM (Simultaneous Localization and Mapping) is a technique used to identify the surroundings and create a 3D model, which is used for autonomous navigation, in association with collision avoidance. Here, I have used a SLAM SF-11 LIDAR (Light Detection and Ranging Sensor) which throws IR waves at particular angles, and records the distance with the current angular position.

    Tethered Unmanned Aerial Vehicle for border patrol

    Another defense project which has endurance as it’s primary feature. The stipulated design does not require range/wireless flying. The requirement is to satisfy very long hours of flight without overheating the avionics. This project is meant border patrolling with online web cameras that provide live video feeds to a central server for constant surveilance.

    Autonomous indoor Navigation for AR Parrot v2.0

    The project concludes that by leveraging the Parrot AR Drone client API through Node.js, designing an Ultrasonic System and understanding the Drone’s Serial port architecture, it was possible to create a Drone capable of computing spatial data to avoid obstacles and navigate indoors autonomously. I decided to make it Open Source so that anyone can build it and hopefully improve the system!

  • real time Human detection and tracking using drones

    Another defense project of which this is a part. The project is classified so the ​ distribution of the whole design is forbidden. But this ​human tracking part is basically detecting and tracking human beings in real time, in case of any intrusion. The drone is supposed to identify the intruder (as a human being) and track to neutralize.

    Facial Recognition using lbpf algorithm in python

    Facial recognition libraries are not available easily online, so I tried doing something on my own- The first step is to find a good database of faces with multiple images for each induvidual. The next step is to detect faces in the database images and use them to train the face recognizer. The last step is to test the face recognizer to recognize faces it was trained for. Achieved accuracy of upto 50%

    Dynamics of Galaxies Using N-Body Simulation

    In physics and astronomy, an N-body simulation is a simulation of a dynamical system of particles, usually under the influence of physical forces, such as gravity. This model helps create a model of the galaxies using N-Body Simulation theory. It therefore, helps astrophysicists in studying the various dynamics, like formation, evolution and disruption.

  • Computer Vision Assisted Automated Drill for Dental Surgeries

    A vision guided bot that can be used for dental drilling and other tooth operations. Once the position is specified, it reaches the particular coordinates and drills at a user specified RPM and torque.

    IoT-based Cloud Health Monitoring System

    The project aims to provide medical assistance to rural population with the help of electronic hardware and cloud platform so that doctors can remotely examine the patients and study the reports.

    retinal imaging and analysis using matlab

    The code detects bright spots in images using basic image processing techniques. Can be used to detect retinal dysfunctions. This project was also done in association with LV Prasad Eye Institute and MIT Media Lab.

Excellence is not a skill.It's an attitude.

Software Skills

Prgramming Languages & Scripts

Packages & Tools

when tired, take rest.but don't stop.

Hardware Skills

Microcontroller Families and DAQ's

SoC's and FPGA's

Problems are not stop signs.They are guidelines.

Wanna Be Friends ?

Address

Unmanned Aerial Vehicles Lab, Aerospace Engineering, Indian Institute of Science, Bangalore, Karnataka, India - 560012

Phone

(+91)-7730903488

 

Email

f2013114@hyderabad.bits-pilani.ac.in

kumaar324@gmail.com