SUMMER PROJECT ROBOTICS CLUB, IIT KANPUR

Transcription

1 SUMMER PROJECT ROBOTICS CLUB, IIT KANPUR 2013 AUTONOMOUS UNDERWATER VEHICLE Mentor - Dhrupal R. Shah Hall-10 Mob: Members Prakhar Jain Hall-5 Mob: Pranav Vyas Hall-3 Mob: Akshay Sunil Masare Hall-2 Mob: Jithesh Kumar Hall-2 Mob: Ajay Bajaj Hall-3 Mob: Deepak Kumar Hall-3 Mob: Veena Ahirwar GH-1

2 CHASIS AND BALLAST MECHANISM 1. We started with the basic model prepared before vacations that had many flaws, it lacked theoretical verification. The model was not accurate, the focus was on getting the project approved.(pic) 2. Soon after we came back from holidays, we decided to focus on the design as it was the most the important part the robot. The initial design had two syringes along the length of the robot. But Jithesh, our team member found it was a better idea to have the syringes in the front and back of the robot as they could then be used to control the pitch of the bot and it would also remove the need of moving counterweights inside the bot. He cited it from some source. 3. Our initial mechanism to pull syringes open was rack and pinion mechanism. We discovered that Winch mechanism which implemented pulley was better. We discarded the previous mechanism because it involved movement of mass which resulted in change of center of gravity and we would have to put movable counterweights to maintain its position. 4. Then we focused on dimensions of the robot and improving the design. We decided to have two different compartments for electronics and mechanical parts of the bot. We checked the approximate dimensions of the various components to get an overview of size of our going to be robot. 5. Our next point of discussion was the material of the body of the robot. We looked upon various plastics and metals. We thought we found something useful in Acrylic. Later we discovered various cons of using it like its brittleness etc. Then we began looking for some metal and soon discovered that aluminum qualified all our requirements. It is anti-corrosive and durable. 6. The next agenda item was deciding the shape of the robot. We wanted the bot to be efficient which obviously required us to be in harmony with nature. We studied a bit about fluids and aerodynamics and to keep the drag force minimum, we decided to have all the edges curved. But it turned out that such a design would be difficult to fabricate. Finally after discussing the pros and cons, we preferred two rectangular boxes, one for electronics and mechanism each. We decided to give a try on aerodynamic shape, later on. 7. Next in line for us was a late night discussion with Robotics Club Coordinator, Abhishek Attal who suggested us to concentrate on making the bot user-friendly by making the two parts of our bot separate yet unified. The idea was to have the two boxes separately waterproofed, which could be disconnected and worked upon separately. 8. Now we decided the placement of racks and for different electronics components and for placing batteries. 9. The design got complicated with each step. By then, we had two different boxes, some serious waterproofing issues and so many different mechanical parts. So, we decided to get the body of our robot fabricated by a Bangalore based company. 10. Now this is something we realized very late. We found that Winch mechanism which involved pulleys and threads was unreliable due to the limitation of the strength of the thread. Now what we are using right now, is something called Acme Screw Driven mechanism. (Pic)

3 MOTORS AND PROPELLORS 1. The initial model during abstract submission was not made after rigorous research. We had an idea that we needed high torque motors for moving the rack and pinion and propellers to counter the drag force of water. 2. Afterwards, on some research we found brushless motors were very common in most of the AUVs and RUVs. But it took no time for us to realize that they were way too costly. So we continued with normal high toque 300 rpm DC motors, at least in this prototype. 3. The Winch mechanism required servo motors capable of making more than 3 full rotations. So we chose this servo: HITEC HS-785HB 3.5 TURN WINCH BOAT SERVO. 4. Finally, we switched on to the Acme screw driving mechanism and that required a motor that could make large no. of rotations simultaneously providing sufficient torque. We tested with 10 rpm, 100 rpm and 200 rpm motors and found 100 rpm motors were most suitable. 5. Large surface area of propellers provide large torque. But at the same time it increases the drag force. So thinner blades are better for higher top speeds while larger, flatter blades are essential for better acceleration. So we decided to have a 4-5 blades propeller with 8-10 cm diameter. Propeller s blades of thrusters need to be rotated in opposite directions to cancel the torque generated by the motors and hence the propellers blade are required to come in a right and left handed set.

4 ELECTRONICS 1. We began with a very basic model. The Robot we were designing was supposed to be manually controlled. It was going to use a single ultrasonic sensor, in the bottom to keep a track of the depth and of course an Arduino board. To control the bot, we discussed the possible wireless communication methods namely Xbee and use of PS2 controller. We were advised to use PS2 controller as it is easy to use. 2. During the time of project discussion, Coordinators advised us to go for an Autonomous Underwater Vehicle. Then began a serious brainstorming session that included discussions, discussions, discussions and yes, a lot of googling. 3. We decide to use Arduino Mega for it has more number of pins and expected to have a lot of sensors in our bot. We had no idea how were we going to control the pitch of our robot. It was then that we discovered a device called Gyroscope which could be used to measure the change in angle. Soon enough, we realized that we needed to know the acceleration and velocity of the bot. This led to the discovery of accelerometer. We then found that a device called IMU integrated both these devices on a single board. So finally we decide to use IMU 6 dof. 4. Next we concentrated on obstacle avoidance. We realized the need of a total of five sonars to keep a track of obstacle in every direction, the bot could move (Our bot could not move backwards). So till then, we had got the idea of a 6 dof IMU and use of 5 sonars. 5. During one of the googling session, we found that we could instead use a 9 dof IMU which incorporated a magnetometer as well. This was necessary for us to keep our bot in synchronization with the ground frame. The 6 dof IMU gave the data but in the bot frame. Magnetometer, however, gave us the deflection of our bot from Earth s Magnetic field. Thus we had an idea of the direction of motion of our bot. So IMU 9 dof was in. 6. When we came to know about IMU 10 dof, we decided to use it as we could then have a track of the pressure inside the bot which is of great significance for an AUV. 7. We had the ultrasonic sensors by now. We programmed it and tested it. The results disappointed us. We got correct reading in air but the sensors failed us when we tested them in water. Sensor always gave null distance. Initially we blamed it upon the air-water interface. We then removed the wire mash and got water completely inside the transmitter and the receiver of the sensor. Still the distance was incorrect. No clue why? 8. So the next task was to search for some useful sensor. We left no stone unturned. We mailed many companies, called some. Finally we got a reply from a company which sent its catalogue of products and some useful links. We realized that we couldn t continue with the idea of using ultrasonic sensors due to budget constraints. We looked for alternatives and on discussion with Atulya Shivamshree (he had made a model of fish in the past), we decided to use IR sensor. We had to compromise on range, but at least we had something.

5 POWER 1. The 4 motors used have a maximum current usage of 7.5 A. So motors total up to use 30 amp max at any instant. 2. Arduino, Motor Driver, IMU and sensors use a maximum of 20mA. 3. So the total power calculations turn out to be 20 Watt. 4. According to these calculations we finally decided to use a single Net Botix 11.1 Volt 5000 mah battery. 5. This would give us between 45 min to 1 hour of continuous bot usage (taking factor of safety 3).

6 FUTURE STRATEGIES 1. Now that we have got our design ready, what is left is the electronics part. As soon as we get the body of the robot with us our team would split up into two. 2. One of the teams would concentrate on turning the body into a functional ROV. Using this initial model we would get a hands on experience of how our mechanical system needs to be operated under different circumstances. This would be really helpful while coding for autonomousity. 3. The other team would be working on coding and electronic aspects like IMU, PID Controller, Filters, Path Planning and Obstacle Avoidance, Sensor feedback, Motor Controlling using PWM.