Robotics

Micromouse - Autonomous Maze solver

This was my first attempt in the field of robotics. Micromouse is an autonomous maze solver robot that is capable of finding the shortest path to the center of an unknown maze. The robot is placed at the farthest cell in the beginning and it calculates the shortest route based on the information from the surroundings.

A two wheeled robot, with a stepper motor in each wheel avoids walls and corners with the help of three IR sensors. The brain of the robot is an ATMega32 microcontroller loaded with floodfill algorithm which finds the shortest path to the center. I used the “Bellmann Floodfill” algorithm to find the shortest route and programmed it using C language in AVR studio. The whole maze is stored in an array of 16x16. Another array stored whether the particular cell is being visited earlier or not. I converted the sensors output voltage into 0 or 1 using ADC conversion.

Main functions in the code:

v Forward(); \\ motors forward

v Left(); \\ motors left

v Right(); \\ motors right

v Turn(); \\ motors turn

v Adc_convert(); \\ converts the sensors output into 0, 1

v Floodfill(); \\ fills the maze with distances

v Update(); \\updates the cell value

v Compare \\compares the sensors value and cell

value and gives desired outputs

Mobile Robotic Hand (3+2 DOF Manipulator)

A mobile pick and place robot with a base of 2DOF and manipulator arm of 3DOF was fabricated using plywood. Most of the manipulators in the industry either follow a predefined set of instructions or are controlled by a human. However, my aim was to make an autonomous manipulator which can detect an object and orient itself accordingly. The biggest challenge in this was to decide the DOF of manipulator because more the number of DOF more is the complexity.

The robot used 3 sensors for avoiding obstacle in movement and 1 sensor for object detection. It used 5 servos for arm rotation and 2 DC geared motors for the mobile base. The sensor for object detection scans any object in its range and updates the ATMEGA32 microcontroller for necessary action. A bump sensor mounted on the wrist notifies the microcontroller when the object is held between the fingers. The payload of this robot is 50g-75g but this can increased by using motors of high torque. It has a maximum arm stretch of 50cm and a gripper range of 10cm. NiMh batteries are used for power supply. Coding for desired robot was done in C using AVR Studio and burnt in microcontroller using FreeISP using a serial port.

I wouldn't say that it was a success as the fabricated prototype lacked a lot of improvisations like object detection using a camera along with a switch to manual control or an RF control. But I developed this on my own in less than 4 months as a part of project course and learnt a lot of things. While working on this, I watched the DARPA grand challenge documentary which was very fascinating and inspiring. I also felt the need of learning Image processing and implementing it. I attended workshops related to Image processing to learn this aspect. However I did not implement it on this bot.

Walking Bipedal Bot

Team: Rahul, Aditya, Sai

As a part of Robotics course, in a team of two, we wanted to develop a robot which did not use much of electronics. So we tried mimicking the human leg movement and designed a functional bidepal robot in a short span of 7 days. We were the only group in the class who were able to demonstrate a functional robot.

The robot consisted of two legs connected to a platform on which there is a weight which balances the robot as it moves. A functional walking robot used 3 DC geared motors – 1 each at the base of the leg and 1 at the top centre to position the balancing weight thus forming a shape of π. In the first stage the robot will shift its weight to one side thereby lifting the other leg. In the second stage, the motor on the robot leg will be powered causing the robot to move about the leg. Finally the weight will be shifted to the centre and the robot will be balanced again.

The body was made of wood and the motors used were DC motors – two motors on the legs were of 45 RPM and the one used to shift the weight was 100RPM. Power was supplied externally to the robot and controlled manually using a remote control having three DPDT switches.

RoboWarrior

Team: Aayush and

The aim was to design a combat robot with intent of defensive strategy. The robot was designed in a team of 4. The fabrication was done by the rest three.

This 22kg robot employed 2 high torque DC motors with torque ~ 40kg.cm. The rest of the body was rested on castor wheels and the chassis was maintained as low as possible to provide minimum ground clearance. We used two sharp rods for damaging the opponent. These rods were welded in the front of bulky Iron wedge. A grinding wheel rotating at 200rpm was used to protect the rear side of the robot.

Google Sites

Report abuse