Human & Robot Interaction Lab. (TaarLab)

Login
A+ A A-

Single- and multi-agent control of mobile unmanned vehicles, including among others, mobile robots, are of paramount contributions to evolution and enhancement of human-robot and robot-robot interactions in TaarLab, from both practical and theoretical points of view. The main goal of our research projects conducted on the domain of mobile robots is "Controlling a mobile robot in order that it could perform a desired task safely and autonomously, whether or not being in communication with others ".


For an experimental purpose, a mobile robot should possess at least some of the basic technical facilities, such as proximity sensor, camera, accelerometer, sound sensor and wireless communication tool. Consequently, in the beginning, TaarLab was equipped with seven e-pucks, for they, despite their diminutive size, report thorough sensory odometry information, and as well, are priced fairly at nearly 900 USD. Afterward, in order to make way for faster implementation of some real-time procedures, such as optimal vision-based obstacle avoidance, TaarLab developed a FPGA-based mobile robot, called MRTQ.

In what follows, some captions of our research projects carried out on the field of mobile robots, using either e-puck or MRTQ, are provided and briefly explained.

Mobile robots for implementing SLAM via data fusion concepts

People Involved
   Arya SabouryAli Ravari

Designing open-source educational robots has a great impact on empowering engineers and those interested in mobile robots. In the last decade, different robots have been designed and built for this purpose. In this project, with the help of sensors available on a 9-axis open-source IMU and a robot based on open-source hardware and control software, a graphical user interface has been implemented to view the condition of motion variables in a real-time manner and combine sensor data. Moreover, by implementing different filter programs, such as low-pass filter, complementary filter and by using KF implementation and deliberation this mobile robot becomes suitable for improving the accuracy of the robot in extracting parameters required for localization, mapping and robot navigation.
Features of the developed robot, the so-called TL-OSR(TaarLab-Open-source robot), can be regarded as: using open-source hardware and software with real-time performance, low price equipment, presenting a solution based on sensor fusion for better accuracy in robot navigation and using different types of sensors. This robot can be used as an suitable open-source platform as a new class of mobile robots for educational and research purposes in mechatronics and robotics.

TL OSR2 TL OSR1

Motion planning and MPC for group of E-Pucks connected by cable

People Involved
   Roya Sabbagh Novin

Mobile robots play an important role in human-robot interaction and they can be regarded as one of the major focused researches conducted in TaarLab. In this context, path planning plays an important role and various methods have been implemented so far, such as artificial potential fields.

However, a reliable real-time path planning algorithm can be much more practical. Thus, convex optimization is used for path planning and a model predictive control (MPC) system generates the optimal control variables. A two wheeled mobile robot system such as E-puck, has two inputs (translational velocity and angular velocity) and three outputs (center positions x, y and heading angle of the mobile robot on two dimensional Cartesian workspace). The controller can track the desired path by using these inputs and feedback of the outputs.

This framework also enables multiple mobile robots with limited ranges of sensing and communication to maneuver and achieve goals safely in dynamic environments. To combine the respective advantages of centralized and de-centralized planning, this framework is based on the concept of centralized planning within dynamic robot networks. As the robots move in their environment, each robot tracks the centralized path plan with respect to its own constrains. In addition, a version of this problem is solved with cables connecting E-puck with each other.

TaarLab is equipped with 7 e-pucks which are being used for practical tests.

NGN

Obstacle avoidance algorithm on mobile robot via vision and locost sensor

People Involved
   Siavash BoroumandArya Saboury ,Ali Ravari, Amir Nasser Khaleghi

With application in the robotics and automation, more and more it becomes necessary the development of applications based on methodologies that facilitate future modifications, updates and enhancements in the original projected system. This project presents a conception of mobile robots using rapid prototyping, distributing the several control actions in growing levels of complexity and computing proposal oriented to embed systems implementation. This kind of controller can be tested on different platform representing the mobile robots using reprogrammable logic components (FPGA). This mobile robot will detect obstacle and also be able to control the speed. Different modules will be Actuators, Sensors, wireless transmission. All this modules will be interfaced using FPGA controller. I would like to construct a mechanically simple robot model, which can measure the distance from obstacle with the aid of sensor and accordingly should able to control the speed of motor.

MRTQ

Vision-based trajectory planning of E-Pucks with obstacle avoidance

People Involved
   Amir Nasser Khaleghi, Siavash Boroumand

The main objective of this study was to evaluate and implementation of mobile robots in unknown environments. The strategy used in this project is the use of image processing techniques and BUG algorithm. In this study, using image processing algorithms robot, target and obstacles are identified, then the angle and distance from the robot to the target is calculated, The robot then rotate, move towards the target. If an obstacle on the way to deal with the obstacles avoidance algorithms passes and continues its path to the target.

mobile5

RoboCup Junior Project

People Involved
  Navid KashiPourya Shahverdi

RoboCup Junior is a project-oriented educational initiative that sponsors local, regional and international robotic events for young students. It is designed to introduce RoboCup to primary and secondary school children, as well as undergraduates who do not have the resources to get involved in the senior leagues yet. The Junior League has three different competitions: Soccer, Dance and Rescue. In the junior rescue league robots identify victims within re-created disaster scenarios, varying in complexity from line-following on a flat surface to negotiating paths through obstacles on uneven terrain. Taarlab started junior projects with a Junior Rescue robot that has made by some talented students from National Organization for Development of Exceptional Talent (NODET) in Iran and other special high schools. The leader of the junior projects is Navid Kashi and this team is supervised by Pourya Shahverdi. The focus of the Junior League lies on education. Taarlab Junior-Rescue team was participated in the 2014 IranOpen Robocup and awarded the 3rd place in this competition, so this team has qualified for the 2014 International Robocup in the Joao Pessoa, Brazil. TaarLab has registered for this international competition, but unfortunately the students couldn't participate in this RoboCup. TaarLab is decided to participate in the next 2015 IranOpen RoboCup.

mobile6

Spherical Mobile Robot and Path Planning Based on Visual Servoing

People Involved
  Rasoul Sadeghian

In recent decade, research about spherical robots as an effective future technology is more expanded. Spherical robot is a non-holonomic system with a lot of applications like environment exploration, security guard or entertainment device. There are various drive mechanisms as Hamster ball, pendulum drive mechanism, and shell transformation.
The mechanism which constructed is a novel drive mechanism for spherical robots. This mechanism is combined of two different types of drive mechanisms. Pendulum and hamster ball mechanisms are combined to a new drive mechanism. In this drive mechanism a servo motor is used to steer the pendulum stick to left and right sides and two DC-encoder motors are used instead of the pendulum bob. Various controllers are used for controlling the robot, but the main controller applied is based on visual servoing. The other controller, which is applied are PID, Fuzzy and MPC. Controlling the spherical robot on an unbalance plate is the future goal of this project.


ballrobot1 ballrobot2

Control a Two-wheeled Self Balancing Robot (TSBL)

People Involved
  Rasoul Sadeghian

In recent decades, many researches have been conducted on urban transport non-pollutant vehicles. The main features of these vehicles which make them attractive for users regarded as no need fossil fuels, need a little space for movement and high ease to use and great maneuverability with independent control of two wheels. There is a great deal of several controllers were used for two-wheeled self-balancing robots. The most applied controllers for this sort of robot are like PID back stepping controller, robust controller, Fuzzy controller and a bunch of various controllers can be used for stabling the two-wheeled self-balancing robot. The self-balance robot (TSBL) it is built in TAAR laboratory is a prototype for implementing the different controllers to find the best controller for robot balancing. The control system used in this robot was based on two decoupled state-space controllers, the first one control, stability about the lateral axis (pitch) and the second one acting on the dynamics around the vertical axis (yaw).


In this project, the PD, PID and Fuzzy controllers are applied to improve the actions of TSBL robot. Some of primary results shown that the PD controller made TSBL robot balance but not stable. The PID controller made TSBL robot balance and stable, but a slight vibration appear on robot body and the Fuzzy controller made the robot stable, balance and omit the body vibration of robot. In future work MPC and hybrid controllers will be implemented in TSBL robot for finding the best controller.

self balance robot

Robot for Obstacle Avoidance and Human Following

People Involved
  Ali Ravari, Arya Saboury

TL-PR is a robot like at home or presents robot. This robot is designed for implementation obstacle avoidance and person following.  At the same time, this robot use open source hardware. In TL-PR, Kinect uses to extract images from the area around the robot. Ultrasonic sensor use for Obstacle avoidance.   Now, the students in TaarLab under the supervision of Dr. M. Tale Masouleh are researched about image processing algorithms, obstacle avoidance, human following and sensor fusion and implement of them on robots.
Features of the developed robot, the so-called TL-OSR (TaarLab-Open-source robot), can be regarded as: using open-source hardware and software with real-time performance, low price equipment, presenting a solution based on sensor fusion for better accuracy in robot navigation and using different types of sensors. This robot can be used as an suitable open-source platform as a new class of mobile robots for educational and research purposes in mechatronics and robotics

mr1

Control image processing for mobile robot for a desired path tracking purpose

People Involved
   Maryam Hadi Koloo

Description: Robot performance improvement and control in diverse industries are very important, especially for the applications that are needed to high accuracy, quick action and high efficiency.
In this research a dynamic base robot that includes two electrical motors and an IR sensor, has been investigated to determine its location through image processing, in a reference coordinate system. it has been assumed a high accurate path tracking and optimum speed for this robot. To control the robot an integrated strategy of Feedback and feedforward has been applied. Fast transient response will be achieved by feedback control via pole placement and by feedforward control, the robot would be tracked the path profile.

e puck

Flexible amphibious mobile robot controlled by teleoperation

People Involved
   Mohammad Safary

Patrol-S is a novel flexible Amphibious robot design and constructed by Mohammad Safari in Taarlab. The unique mechanical structure design of this robot makes it suitable motion for a wide range of environments, such as rocky, sandy and uneven. The half-circular-legs are used on the external shell of wheel with the aim of helping obstacles climbing. The material for the construction of this part has been selected upon performing done different practical analyzes. The dynamic simulation of the proposed robot has been done in MSC.ADAMS software. In particular, the decisions about hardness and width of robot parts, optimal range of spring stiffness factor and the required torque to drive the robot with suitable speed is done by simulation based on extensive analytical and scientific discussions. According to the desired task envisaged for Patrol-S, such as patrolling, environmental monitoring, environment identification, search and rescue applications, user teleoperation is also considered. In order to detect the height of obstacle, several Ultrasonic sensors are installed in the front of robot. The user could assign this phase of movement to the automatic navigation system to the end of improving the performance in making the decision of how the obstacle should be avoided. The robot navigation using sensors data is based on a controller which used Fuzzy Logic.

TL OSR2 TL OSR1

Rescue Robot

RescueRobot

e-puck

e-puck Robot

 

MRTQ: 3rd Place in IranOpen2013

MR 02

 

MRTQ

MRTQ Robot

 

 

TaarLab 2014, All Rights Reserved.
Designed, Developed and Powered by: Payam Ghasemi & Nima Karbasi

Login