Surface Water Waste Collection Robot for Community Impact Under EPICS-in-IEEE

As a part of EPICS, I got the chance to conduct a study on the existing Waste collecting ASV's and design a fully autonomous and durable design with minimum cost on maintenance for IEEE SIGHT CEC.

Also, all the 3d printed parts shown in this entire project is designed from scratch and printed by me under DNA.rc . All rights reserved

Frame Essentials (f):

1. 12 and 30mm OD Hollow 1000mm Carbon fiber rods : Buy High quality and strong 3K Roll-wrapped Carbon Fibre Tube Hollow
2. 1 inch SS hollow rode 15m
3. Custom 3d printed joints.
4. Brass Inserts m5, m6 : M5X12MM Brass Insert.

Thruster (t):

1. 2 X 5X8 3d printed propellor ( made for initial prototype).
2. 2 X 3550 underwater thrusters with 80mm props
3. 2 X 3660 SURPASS waterproof BLDC : Buy SURPASS HOBBY- 3660 ROCKET sensoreless waterproof motor at best price.
4. 3 mm 1 meter square Sheet Metal plate.
5. 3d printed thruster mount.

Electronics (e):

1. Flight Controller: Pixhawk PX4 with power supply (not used) : Buy Holybro Pixhawk 4(aluminum case)&PM07 14S Set
2. Raspberry pi 5 : Official Raspberry Pi 5 8GB Starter
3. 2 X 80 A bidirectional ESC(Custom Made).
4. Solar charge controller Buy 60A Intelligent LCD Solar Controller Online | Robu.in
5. 4 X 50A ESC
6. 135,55W mono perk Solar panel, with 100Ah LIFePo4 inverter.
7. 2 X 15 Ah 3C Lithium iron phosphate Battery (Custom Made).

Insight 1

Using an upper mounted gearmotor for actuation of an open and closable front flap.

Insight 2

Using a simple fit in place basket with partial enclosure which helps in an easy pull up mechanism to collect waste.

Insight 3

As mentioned in the structural design, we were only able to collect 40 - 60 L of waste so came up with an idea of an expanding back net helping to increase the active surface area there by increasing the potential volume of waste that can be collected.

Above is a 3d printed 1:10 scale of our project I made to present my design to the IEEE team.

Frame Design

02-02-2024: The main frame of the ASV is made of (f).1 with dimension 1000mmX1000mmX420mm and uses (f).2 i.e 3d printed 3- and 4-way joints.

Total Volume of 440L, considering Buoyancy and floating nature of plastic the ASV is expected to collect 40 to 60 L of waste in a single run.

19-12-2024: After that the model was scaled to 80 percent and continued to be used till now.

Propulsion System

Thruster Mounts: In the current stage the thruster is attached directly to the main frame on each side with 5 mm bolts

Steering Mechanism: Uses differential thrust

Hull design

02-12-2024: Shape: As we are using a catamaran model ASV with 2 identical and symmetrical modular hulls.

Material Selection: At Stage 2 prototyping, as of now we are using a 1400mmX20mmX40mm 3d printed hull design (not tested/simulated) with epoxy to water seal the hull.

For Stage 1 prototype we used a 6-inch PVC pipe

03-03-2025: Now a fully 3d printed 160cmX250cmX250cm hull printed out of pla+ is used in the Stage 2 model

Structural Analysis

As of now I just kickstarted Ansys fluent for the hull and structural dynamics Sim. Also, after stage 2 completion the tested files and sim data will be open-sourced :) <3.

21-05-2025: As of date not much progress is done on sim side as I backed to Og role of design and manufacturing lead in the project.

For thrusters we initially used a 2836 750kv BLDC and for that used a 5inch diameter and 8inch pitch tri blade propellor(t).1.

02-12-24: Currently we are experimenting with newer propellors, and suggestion are appreciated.

Doubts:

1. Material Selection for Hull: Fiberglass/carbon fiber or simple 3d printed hull with epoxy to water seal, which would be better for our use case.
2. Propellor Design: Considering the size of the ASV what the preferred pitch and diameter of the propellor?
3. Thruster Mounting Location: In the miniature design also, we were not able finalize the position for thruster that would result in better efficiency for the ASV.

Conclusion:

1. Material Selection for Hull: PLA+ is fine as its ease to print with clean surface even compared with PETG, as ABS needed enclosure fumes where too much and so the print failures and strength test failure for thin parts.
2. Propellor Design: Efficiency is not only about pitch and diameter—refer to the rctestflight challenge below for a practical approach. I Found the Most Efficient Propeller Design - Competition Ep. 3 - YouTube
3. Thruster Mounting Location: Thruster position alone doesn’t determine efficiency. Overall motor capability, operating throttle, and ESC response characteristics have a larger impact.

Stage 1 prototype was used to experiment with different motors, propellors, speed and buoyancy.

Team Members (prototype making):

1. Jerrin Shibu (me, back left), 4th year CEC
2. Vignesh J R (front right), 3rd year CEC
3. Vijay Satheesh, 3rd year CEC
4. Yohan CM (front left) ,4th year CEC
5. Bhavesh Sanjay (camera guy <3), 3rd year CEC
6. Jeffin Shibu (back right), my mentor & postgraduate, IITM

Observations

1. 2836 motors are not sufficient
2. FDM 3d printed propellors gets damaged easily
3. As we got to test the robot at evening lack of lighting really made us have a hard time.

Discarded Ideas:

1. Discarded Thruster Design V1

Updates from Stage 1 Observation:

1. Added onboard lights to illuminate the inside tray
2. Switched to above prebuilt (t).2 thrusters
3. Added a 55W solar panel as per teammates suggestion and a portables 100Ah solar powered inverter was also built by the electronics team.

Modular Hull Made to Stack up distance , each module having dimensions 250X250X250mm.

03-03-25: Currrently Testing out the 3d printer settings with different materials and PLA composites, including PLA LW, PLA + , ASA and PETG.

Currently Using .8mm hardened steel nozzle,and 12 mm Carbon Fiber Rods to connect the modules

Finally revised the design by taking inspirations from other fellow creators.

3d printed the hull 2 hulls, one with mixed infills for strength validation(right) and other with 10 - 15 percent infill to test weight reduction(left) with a total of over 40+ functional parts and 400+ print hours at 400mms and .48-.4-layer height.

Initally assembled the hulls with duct tape and friction fitted 12 mm hollow CF rods

The partially assembled hull was tested on nearby canal.

Observation:

1. Zero noticeable leaks.
2. As the wall thickness is only 1.6mm a strong impact on sides and add unnoticeable layer tear, so a coat of epoxy or fiberglass would be preferred.
3. As am 70kg and its total volume is above 140L, Custom hull simply increases the look and mechanical properties anyways.

Assembled with clear silicon and m5 & m6 bolts and left should leave the silicon to dry up for min 3 days to a week.

Also, it's better to leave it upright and push down gentle from above after the assembly.

Assembled the Hull to the improvised old frame(800cmX800cmX32cm) with custom multipurpose handle and all other 3d printed parts including custom waterproof containers, frame strengtheners and mounts.

Observations:

1. When you fill silicon while assembling add it much more than you feel needed else minor air gaps can potentially lead water leaks and was observed in hull1 as we filled hull 1 with enough silicon and 2 with much more then needed.
2. Plastic Props and not build for tough environment.

Got noticed by some media at end of stage 2

Meanwhile will be adding

1. Automated Tray and Flap system.
2. As the prebuilt props aren't tough enough, new propellors are resin printed in ABS like resin.
3. Earlier, I designed the tray to be removed and refitted vertically, but after integrating the onboard solar panel, the tray is now designed to be removed horizontally instead which can be quite uncomfortable to the user.

“About 8 million tons of plastic enter the ocean every year, and I can’t do anything about that. But maybe, with a little bit of creativity, I could build a 3D-printed boat that anyone could make and use to skim plastic out of water all over the world.” – DrewBuiltStuff

In this project, I was responsible for the design and manufacturing. Initially, it was meant to be an autonomous system, but over time its purpose shifted, and it ended up being mostly RC-operated, as the software leads lacked the experience to work with industry-standard peripherals and funds were limited.

With my previous autonomous builds as a foundation, I am now developing a fully autonomous MK2 version from the ground up, independently, under DNA.rc.

By Stage 3 testing, the project is expected to conclude with the release of all revised design files.