2 of 13

How did we come up with the design

We knew we wanted to do treads due to their amazing performance off road but realized building a tread system would be complicated so we opted for a pre built chassis. This chassis also had motors geared for torque which was important
We knew arduino would be complicated but it also was highly customizable and adaptable to our needs
Wireless was a must so we found out about the NRF24L01, an easy and impressive feature rich 2.4 GHz wireless card which was better than IR or an rc radio communicator
We wanted to have an articulating arm for the payload part and a gripper at the end for maximum payload flexibility
The controller also had to be an arduino and we decided on one joystick to control the rover and the other plus a few additional buttons to control the arm. The whole casing was 3D modeled and then printed custom for this project
We decided on smaller, lighter 9g servos for the arm but that ended up being a bit of a bad decision
We wanted to be able to efficiently do the light part so we built a pole, with the lights at the right heights
The arm also provided great flexibility we decided, it could work on the payload drop off, payload pickup, and the mechanical arm part
We decided on 3.7 Li-ion batteries because they could deliver the 3A maximum motor draw, power the transmitter and everything else for a considerable amount of time due to the 2000 mAh capacity. We wired two in series for a measured maximum voltage of 8.5V

3 of 13

Testing

At first the biggest issue was getting the NRF24L01s to work. The breakout boards made this considerably easier because we could use a 5V power supply from the motor controller even though the NRF24L01 used 3.3V. We eventually found an example sketch for the NRF24L01 and heavily modified it for our use.
Another issue was the L298N motor controller. The first one we got was dead on arrival so we ordered a new one. In the meantime we were nearing the original launch date so we opted for a relay shield which eventually came in useful for the LDR challenge and not the motors. The working L298N gave us the ability to use speed control and eventually worked with the arduino
The relay shield was used to control the light and we had an error transmitting the variable that controlled that but it mysteriously fixed itself
One issue we never fixed was that if the rover lost signal for some reason, it began to drive full speed backwards. We had a plan to fix this but never got time to
When we tested the treads outside, we discovered that the high torque motors and treads did great over large obstacles, One issue though was that it slipped when on an edge so we added high-tech beads of hot glue to add a ledge that helped get over many more objects
We also discovered that the servos were far too weak to lift the full articulating arm we wanted so we just secured the 3d designed and printed arm directly to the rover

Testing (cont.)

Another issue we encountered was that our robotic arm did not reach the ground which would prove to be an issue for the smaller items. So we added small metal rods with hot glue at the end(for grip) to reach the ground
Another traction issue we found out about during testing was the tendency of the rover to flip on it’s back if the object was too tall. We solved this with the green wheels seen in the back of the rover. This is a “wheelie-bar” which prevents it from flipping over
As for the controller, when we tried to affix the lid, we realized that the 3d model had not taken into account the height of the jumper cables. To fix this a lighter and a metal rod was used as a quick fix to mold an indent
The object that was being manipulated was hard to see and we didn’t know what view we were going to add to we added an old Android phone as an FPV system

The NRF24L01 and breakout board

The L298N Motor controller

Our LED mounting solution

The manipulator arm

The three batteries

The arduino mega and relay shield. Also shown is the traction hot blue beads

The final rover and its wheelie-bar

The final controller

Testing! (Pictures)

Testing! (More Pictures)

Launch day!

We did extremely well on the obstacle part, getting the full 50 possible points in 48 seconds
Our tower solution for the light sensor did very well and we got all of the points
The mechanical switch was a little more difficult because our robotic arm had motors that weren't powerful enough so they overdrew the batteries and the rover stalled a few times
This same issue surfaced in the payload retrieval and drop off and the lack of different degrees of freedom made it hard to pick up objects
The arms movement was also not accurate enough in some cases so it was difficult to line the arm up to pick up small objects
One major bug we encountered was that a couple of times the rover got stuck repeating the previous command which proved for a few point deductions
Our last minute FPV system proved to not be very useful

Changes

The 9g servos were far to weak. 55g gram servos would have been far better and if we made it move slower, picking up objects would have been far easier
More time to program would allow us to fix the signal bugs
More axis of movement on the arm would be an extremely helpful addition

Lessons Learned