Using Flow Meters to measure Robot Velocity

[Kolin Szeto]

Teams,

We are trying to find an alternative to the DVL, to track the position of the robot in the pool. Our team has started to investigate using water flow sensors in multiple places on the sub combined with the depth sensor and on board imu in the pixhawk. This will allow us to theoretically track our position in the x, y, and z axis and will attempt to combat water current, which may spin each sensor at a different speed and therefore artificially change the location of the robot. We have not seen this approach taken in any other team's technical paper or website. What are your opinions of the viability of this approach/any pointers/has any team attempted this before? You can take a look at the amount of space that each of the flow sensors take up as we determined a check valve would be needed to allow for a single direction of flow of the water to be able to use the flow sensors in their stock configuration. We have not figured out a clean method (excluding epoxy) of water proofing the flow sensors without entirely encasing it in an enclosure. Let us know your thoughts.

[Daveman]

I don't think I've seen that approach either. Interesting.

If you were just drifting with the current (not that we have a current at TRANSDEC), would you sense it? Since the relative motion between your vehicle and the water flow is zero.

Which parts do you need to water proof for the flow sensor?

[Ledge]

Carl Hayden High School, Falcon Robotics did some testing with this method, Optical Flow. This was before they acquired a DVL. They were using one downward facing camera linked to software that was also linked to depth sensor to calibrate flow with depth from bottom. It does work in keeping track of direction and distance, however the clarity of the water is very important, if you can't see the details of the Transdec floor, such at the algae patches and such, then it will not work or will be spotty at best. This was at least their finding when they attempted it. If clarity is bad then you will need to be closer to the bottom to make it work. You may want to think into getting one of these https://waterlinked.com/dvl/
DVL A50
$5 990.00 – $6 290.0

It sounds like a lot but if you can raise the funds it will be very well worth it! This DVL is going to be a game changer like the Blue Robotics thrusters were a few years back

[Kolin Szeto]

Daveman We are trying to use the built in IMU on the Pixhawk as a failsafe for tracking the movement of the sub if the flow sensors don't notice if the vehicle is drifting. Alternatively we could attempt to use the built in position hold in the program from Blue ROV. But it is a point we will try to dig a little bit deeper. Because the flow sensor is effectively a hall effect sensor packaged with a magnet on the propeller, for waterproofing the sensor, we would add a little bit of epoxy to point where wires come out of the sensor to attempt to block water leaking into the sensor board. The other option we had was placing a flow sensor in an enclosure.


Ledge Thank you for sharing out your past's team experience with navigation with the respective Optical flow sensor. My team has been looking into the DVL you suggested. Thank you for the reference.