Clearbot - Open Ocean Engineering 

Over the past year, I've been working for a team based out of Hong Kong on Clearbot - an unmanned surface vehicle (USV) that uses artificial intelligence to clean up ocean plastic and debris. 

To me, Clearbot is somewhat similar to WALL-E. Clearbot is fully autonomous and "eats" garbage using a conveyor belt. An incredible fact is one robot can remove up to 1 ton (!) of trash per day. Our company hopes to eventually have a fleet of robots that patrol lakes, rivers, and oceans to eliminate pollution and restore our balance with water. 

Below I'll explain what I do at Clearbot and a few of the problems I've solved so far.

Automated Docking System (2022) 

After joining Clearbot, I talked with our Mechanical Engineering Lead, George Jolly, and the CEO of Open Ocean Engineering, Sidhant Gupta, about what I could do to make a meaningful difference at the company. After thinking for a bit, they told me about a project they had shelved for a long time because there was no time to complete it, and they were wondering if I wanted to take it on. While Clearbot is fully autonomous, clients had been inquiring about a docking system that would allow Clearbot to function all day without any human intervention.

​As currently constructed, the robot needed an operator twice a day - once to release Clearbot from the dock and once to tie it back up again to charge. George and Sidhant thought an automated docking system would be challenging, as there are very few products to compare it to. I told them I was ready to tackle the problem and I've included the results below. 

This is my full simulation of the docking sequence, modelled in SOLIDWORKS and edited in Davinci Resolve. On this page, I will explain my design choices and how I created this. 

Stage 1 - Concept Drawings

​The first step for this project was figuring out what dock configuration worked best for Clearbot. George and Sidhant explained that the dock needed to generate power with no external input, it needed to use modular pontoon blocks and costs had to be minimized. While I’m not the best artist, I started off by drawing out 3 design concepts and pitched the pros and cons to the team.

After considering these choices we ended up choosing the 1st option and I drafted a more detailed drawing. At this point the list of design questions I had started getting longer and longer. How would we stabilize Clearbot? How would we secure the robot? What is the amount of error we could have? 

Stage 2 - Preliminary CAD Draft

​Once we were satisfied with the concept drawing, I made a detailed CAD drawing of the system. All of the CAD parts needed to be made by hand, which made the process complex and time-consuming.

Overall, this CAD drawing gave us a way to visualize the dock and understand the path towards a successful product. I created a short animation of the automated docking as well. 

Stage 3 - Design Revisions

​While George and Sidhant were very pleased with the concept drawing, there were a few major problems that we discovered once we got a chance to see the simulation. The first was the dock was too expensive, and I needed to use far fewer modular blocks (grey) to minimize costs. Another problem was I found out that Clearbot can not move backwards! Our current version of the robot only uses forward propulsion, so a major design revision was needed. Lastly, we needed a way to secure the bot so it would not move laterally. Our idea was to use a method similar to slowing down a coaster in a Log Flume Ride. While this may seem radical, we thought the mechanics behind the ride could be applied to the dock. 

Stage 4  - Final CAD

​After a few iterations, I was able to complete the final dock and simulate Clearbot’s entry and exit into the system. The main design change here is the dock almost became a throughway. This allowed Clearbot to enter and exit without having to go backwards. I also added rollers (like the log flume ride), guide boards and electronics. 

Here are a few pictures and the simulation I created.

Stage 5 - Construction (ongoing)

​Parts have been ordered for dock construction. I will update this website with a picture of the finished product when it is done!

---

Automated Charging System (2022)

​While designing the dock, one objective I did not touch on was autonomous charging. This was a complex task and George and Sidhant wanted me to complete this once we had a full picture of what the docking system looked like. So naturally, after completing the dock, I started dissecting the complex challenge of charging autonomously.

Stage 1 - Charging Type

​When I started thinking about this problem, I immediately thought about how I charged my cell phone. Could Clearbot be charged using induction? I researched this, talked to robotics specialists, and set up meetings with Powermat.

​However, this was not a possibility due to the battery type of Clearbot and wattage limitations for inductive charging. George and Sidhant told me to revisit this after a year because inductive charging technology will improve and changes will be made to Clearbot’s battery type. However, at this point it was not a possibility, and we needed to use the traditional method of “plugging in” to charge Clearbot. 

Stage 2 - Initial CAD

​When looking at the bare bones of this problem, we knew Clearbot needed the ability to plug in and charge without any human intervention. For this, I thought using an actuator with a plug on the end would work well because it was a simple way to start off.

​Knowing my design would need to be heavily refined, having this minimalist concept meant there were fewer places to go wrong later. 

Stage 3 - Design Revisions (ongoing)

​I am currently completing a redesign for this charging mechanism. I am looking into blade connectors using an electromagnet and actuator.