Line Following Robot Project

Posted on 2:40 AM by ersinaytac

I was going to write the components of a line following robot part by part, but now i just wanna write what me and my buddy did. About 6 months ago we got this project in our hands, and we didn't now anything really. Thanks to internet we read way too many stuff and watched videos to understand what we'r tryin to do. National line following robot competitions was our target. A mechanical engineering student and a computer engineering student with pretty much no knowledge of electricity describes us perfectly.

What i did was to build a visual prototype out of a paperboard to help us visualize. First idea was making a round chassis out of fiberglass 2 motors in the middle and reach the dynamic equilibrium somehow. After talks with experienced mechanical engineers they convinced me that although i can reach static equilibrium, it is impossible to reach dynamic equilibrium. We choose to use Arduino for micro processor, Ardumoto for motor driver, and 3 Pololu QTR 1RC for reflectance sensors in our robot.

05242010(003)

With a radical decision I bought 2 very cheap toy cars for 10TL (approx. $6) each and put all electronic parts on it. That was our first prototype robot. The problem with that toy car is the steering mechanism consists of a dc motor and spring to correct the steering and a steering rod. The motor turns the steering all the way left or right and the spring turns the steering straight. Also our a cheap toy cars dc motor on the back has a stall current more than 2A which is over the limit of our motor driver Ardumoto. To control the speed, the dc motor is not under continues current. To power the system we used a power adaptor at this time.

IMG_1148

Then we upgraded our car to a slightly better toy car which has differential in the back. Since the steering mechanism was same the result was not fulfilling, also the back motor is drawing too much current. We kept using power adaptor for power.

IMG_1173

The first line following robot was on a HPI semi-pro RC car. The steering mechanism was connected to a servo motor which is perfect for smooth steering. The motor on the back draws way too much current than Ardumoto can drive. So we bought basic ESC (electronic speed control) to drive the motor. The turning diameter of the car was 73 cm which is large for a line following competition. The wheelbase of a car is directly proportional to its turning diameter. With a help of a Dremel, shortened the car by 10 cm and mount the front side (wheels, servo, springs..etc) 10 cm back. After this the turning diameter of the car was 40 cm, which is acceptable for a line following contest. And we changed the 3 individual Pololu sensors to a 8 sensors array. Which is also a significant improvement in the steering. In a national competition this robot finished in 30th place out of 180 robots. We bought 4 very light (one of them weighs 11 g) Li-Po batteries with 3.7V and 350 mAh and connect them in series and produce 7.4V and 350 mAh. One 7.4V set is connected to the motor and esc, and the other 7.4V set was powering the rest of the system.

photo

We decided that the differential drive will give us better results, and built another prototype with 2 dc motors connected to a plastic tank tracks. The 2 dc motors we used on that prototype was Tamiya's double motor gear box. One can assemble the gear box with two different motor ratios. One is 203:1 65 rpm torque configuration and the other one is 58:1 230 rpm speed configuration. Somehow at speed configuration, the two motors did not work simultaneously, as a result the robot didn't go straight well. At torque configuration the motors work identically but the speed is no where near what we are looking for. Also the tracks were causing too much friction. We used Ardumoto to drive this motors. The stall current is on the limits of Ardumoto that's why this prototype didn't come in life. We used the same 7.4V and 350 mAh batteries just one set to power whole of the system.

IMG_1417

For the second line following robot, we saw that the Tamiya double motor gear box and tank tracks were a bad idea thats why we bought 2 micro gear dc motors which has real small stall current and delivers 1000 rpm in 6V and connect them to plain wheels. Ardumoto came in o the scene again to drive these two little dc motors. A light weight rectangular plastic with holes used for the chassis. Motors were mounted in the back top, so we got closer to the surface and the sensors were on the front bottom. With our small design, the robot can make every possible turn even 90 degree ones. This robot finished 11th out of 110 robots in another national competition. We used the same 7.4V and 350 mAh batteries for whole of the system.

IMG_1441

The third and the last line following was pretty much the same as the second one. A really light small plastic board is used for chassis. All unnecessary parts were cut out. What we have at the end is just a frame. Same micro gear dc motors are used and mounted on the back top. And the sensors are front bottom. Although the plastic part was rigid, with the weight of both motors and the batter, which is mounted right in between motors, caused the back part bend and the wheels started to spin. After mounting a little thicker part to the back, the back part become straight and the problem is solved. The robot weighs 120g with batteries. We used the same 7.4V and 350 mAh batteries. If we supply the system with higher voltages the car can go faster. This robot finished 9th out of 400 robots in another national competition.

4609096396_81577aa0c1_b

Here is a short video of the last line following robot we built, optimization before a competition.