Remotely Operated Vehicle (ROV)


This is the final writeup for the MIT class 2.00A FUNdaMENTALS of Engineering Design. This class is offered by the Department of Mechanical Engineering to first-year engineering students.


We formulated and completed space/earth/ocean exploration-based design projects with weekly milestones. It introduces core engineering themes, principles, and modes of thinking, and includes exercises in written and oral communication and team building. Specialized learning modules enable teams to focus on the knowledge required to complete their projects, such as machine elements, electronics, design process, visualization and communication.

The following is a summary of our design project: how we decided on a design, how we evaluated it, the building process and device’s performance.


The goal is to build an Remotely Operated Vehicle (ROV) to score during the final contest.


Each team has 15 minutes to mine objects from open cubical and cylindrical caves. After the start, team members are only allowed to touch their vehicle when it is at arms reach on the pool. They may not tug on the cable to retrieve their vehicle; it must return to the wall on its own. Teams will not be able to see into the cave and will have to rely in the vehicle’s camera. The specific objects’ set up inside their storage caves will not be known.

Lower cave will contain:
– 15 Acrylic spheres 1 inch diameter
– 15 PVC pipes 2 inches long

Upper cylindrical cave will contain:
– 15 Vertical standing capped cylinders: 2 inches long sand-filled PVC pipes with end capes


Here, I detail the characteristics of some of our first designs and their contributions to the final design.


First I considered two one-way doors that did not need additional motors, but whose doors were actuated by the contact forces from approaching the objects to be collected. Circular springs would open the doors inwards. To be successful, a spring constant had to be low enough for the scoring objects to get in but high enough to not be opened by normal movement around the pool.

Assuming a maximum horizontal velocity of 0.38 m/s (velocity calculations are explained in the analysis section), and a blade area of 4.27x10^{-3} m^{3}. The drag force is then F_{D}=0.308N.

As a result, spring mechanism on the one-way doors needs to be \approx12N/m. With expected static friction between the blades and objects of \approx0.6N, this design was not reliable for implementation.

My second design used two motors to rotate inwards two grabbing arms and intercalate.  Assuming grabbing arms 6 inches wide and 3 inches tall and a an angular frequency = 6.4 rad/s, we obtain that a \tau = 0.159N/m

However, this design requires motor precision. Given our constraints, we can not assume that both motors will rotate at exactly the same angular frequency.

My last design was an elaborate grabbing vertical arm to lift the objects to a bucket on top of the object. Calculations showed a torque too high for implementation.


Two of my designs consisted of a bucket or bag system. That is a very simple storing design that would keep the retrieved objects inside the ROV until it comes back to the surface of the pool. My third design was a mechanism that would make a barrier in the lower part of the ROV goes up and down while intercalating grabbing arms and opening and closing the gate. All of my designs used flexible plastic net as the main material, since it was available in the lab, it was light and had many holes that would decrease its contribution to the total drag. This was the final material we used for the ROV.


We evaluated our designs under the following functional requirements:

SIMPLICITY IN DESIGN AND IMPLEMENTATION: Decreases manufacturing time which allows for more time in testing and debugging.

ADAPTABILITY: Must work inside both cylindrical and planar floor shapes. ROV must be easy to modify if needed.

CONSISTENCY: Consistent performance is a better predictor of success than sophisticated design.

DURABILITY: ROV must withstand hours of testing and the short competition.

SIZE: Comply with game’s size rules but not too small as to make implementation more difficult.

COLLECTION CAPACITY: The ability to collect as much objects in the shortest possible time.


This was our decision matrix. We chose to use a Dredger collection mechanism and a Bucket for storage.




The frame is a simple PVC cubic frame.


Our preliminary designs with moving arms and blades were not simple to implement, a requirement of our design parameters. Then, we came up with the idea of using plastic cylinders inserted on a PVC pipe , like bristles. We decided to attach the cylinders using a spiral pattern to maintain a continuous influx of retrieved objects.


The torque required to spin the dredger underwater is given by:

    \[ \tau = \frac{\omega ^{2}\rho B^{4}a}{8} + \frac{B}{2} = 0.023 Nm \]

\rho=1000 kg/m^{3} : density of water
\omega=26rad/s : angular frequency at maximum efficiency
A=12.5cm : combined width of bristles
B=3cm : length of bristles
drag coefficient for water = 1.

However, motor specifications state a maximum power output of 8.02 W. Assuming an efficiency of 72%, the motor can output 5.77 W. This would not provide enough torque, so we added a 100:1 gearbox to the setup, which in turn decreases angular frequency decreased to a maximum of 6.4 rad/s (~1 revolution/s). The resulting torque is then:

    \[ \tau = 0.91 Nm \]


Our simplest collection mechanism was to use a bag with the same plastic net we saw in the lab. That is, a plastic net with holes that would decrease the drag forces by allowing water pass through it. We decided to use an enclosed bag that would surround the lower chamber of the vehicle. That way, if the vehicle accidentally flipped while navigating then the objects would be kept inside.


First, we considered placing two motors on the sides for turning. Our objective was to minimize the number of motors while maximizing the propulsion power. We were given four motors so we planed to dedicate two or three motors for propulsion and one to rotate the vehicle for orientation. However, The decision of using two motors was first motivated by the desire to keep one more motor in reserve. However, we found out that only two the propellers (9 inches wide) would fit on a rod (12 inches long).

Underwater, the ROV is affected by Drag forces that oppose its movement.

    \[ F_{D} = \frac{1}{2}\rho \upsilon ^{2}C_{D}A \]

Two thrusters available for propulsion in any direction provide 8N of Force. This means that the ROV can move at 0.38 m/s when empty. At maximum storage capacity, 6.45 N will be available for vertical propulsion, reaching a maximum vertical speed of 0.17 m/s

By approximating retrieved objects to rectangular shapes, we calculated the total volume of the 45 items in 1028cm^{3}. The total volume of our bucket is 5400cm^{3}, so it could theoretically hold all the objects stacked up. We planned for several legs of trip to retrieve as many objects as possible.

From here, we can calculate the amount of Power and Current required to run each leg of trip. We obtained the following results:
Velocity = distance / time
Drag Force = 80 v^{2}
Drag Power = Drag force x Velocity
Acceleration Power = Mass x Acceleration x Velocity
Gravitational Power = (9.8 x Apparent Mass x Height) / Height
Total Power = Drag Power + Acceleration Power + Gravitational Power
Current = Total Power / Voltage


The first step was to learn to use some of the shop equipment (drills and tools). From then, we started building the frame. Building the cubic frame took half a lab session.


The motors were waterproofed by enclosing them into metal cans alongside their gearboxes. The brush was built using three-inche plastic cylinders to be inserted on the PVC pipe with drilled holes. We added a plastic net bag below the frame and provided some structural support with springs. A spring was located into a screw that was split in two parts and had a separation distance in the middle. That way, the spring could stretch in the desired direction and only up to the pre-established separation distance.

To maximize the efficiency of the LED’s illumination we needed to be able to adjust the angle between the LED board and the camera. We used two elbows and one PVC tee resulting in three degrees of freedom for movement.


Our goal in building the control box was to set up switches for intuitive and ergonomic operation. Switches that needed to be always on during operation (the LED switches) were placed on the box’s lateral plane to avoid accidental pressing.


The ROV was tested on a tow tank and a pool.

7.1. Tow Tank Testing

First, we tested the ROV on a tow tank to check for neutral buoyancy and added floats from empirical observations. We prepared two large PVC floats, two small plastic floats and attached them to the ROV. The ROV sank. We then added one large PVC pipes parallel to the front face to compensate for the brush’s mass and We added two plastic floats at the sides of the ROV toward the back. The ROV seemed to be neutral buoyant.

The day before the competition, the rod rotator motor did not perform reliably. It took several tries to start and had a delay before moving. Later, the internal gearbox failed. It is possible that the gearbox could not handle the increased torque underwater. We also noticed that the motor shaft was not properly engaging the next stages. Regardless, replacement was urgent. A windshield wiper Ford motor was adapted to an aluminum shaft to replace the broken parts.

7.2. Pool Test – Competition Day

Competition day was our first test on the pool. First, the ROV performed adequately. Then, suddenly, the windshield wiper motor failed. We got the ROV out of the pool and evaluated the damage. Our custom-made aluminum shaft was split in two. Without the motor to rotate the two main thrusters, we could not move the ROV in more than one direction.

We added a motor pointing in the vertical direction. However, this modified the ROV’s center of mass and instead of moving in straight lines, the vehicle would only spin. With little time to spare, we looked around the site and found a Poland Springs water bottle that could provide extra flotation with the aid of some duct tape and zip ties.



For the first two minutes of the competition, we attempted to get inside the cylindrical chamber to retrieve the more valuable objects. However, with only one motor on the vertical direction instead of two, we could not navigate well.

We resorted to the safer strategy of going to the lower chamber. However, the springs on the collection net made the ROV bounce up every time we reached the bottom of the pool. With only half the vertical thrust we needed, the motors could not compensate for this bouncing.


In retrospective, we needed more weight in the back for the ROV to reach. Since buoyancy forces are proportional to the height of column of water above the ROV,  the deeper it was, the more force it would perceive.

We did, however, collect two cylinders but the ROV dropped them while coming out of the pool. We were granted half-credit, for a total score of 5. Towards the end of the ascent, the vertical thrust motor failed as well. We retrieved the ROV manually. This is the ROVs after multiple repairs and the competition.