I’ve started riding my bike a lot recently and I’m struck by how much trash is on the side of the highway. Additionally, in California we have a lot of dry grass during the summer that causes fires which doesn’t seem to be well managed.
You see this problem the most on a freeway or a highway due either to litter or just collected from other areas in a city.
So, I had an idea to create a Trash Collector Robots that will vacuum up litter and cut dry grass or plants.
At first we would need to make it remote controlled. This would allow us to manage the robot a little better. Given that the primary use case is near a highway a risk I could see is that the robot if completely automated might accidently go into the street and get damaged or possibly cause an accident.
Additionally, if we have a blade at the bottom of the unit or some other device that is cutting this could be a safety risk to anyone that is in the area. Therefore, once this device is built then when we implement we need to fence off an area that is under work. Another idea could be to shut off the device if it goes outside of the literal fence. That may be a way to put safety measures in place if we decide to automate.
The opportunity here is to adopt a highway to generate revenue. We could also get a contract with the government to take this idea further. The benefit here is that it can save a lot of time, save money, and make our environment cleaner.
What are the parts of this robot?
Wheels – We want the wheels to move independently.
Axle – On a caster wheel there is already a fork that holds the wheel. The motor will need to be attached to the fork and the motor will then need to be attached to the wheel axle to turn the wheel.
Motors – We need four motors that are connected to a fork on a caster wheel. Additionally, there will need to be four other motors that will turn the wheels independently.
Frame – The frame will attach the motors and wheels.
Battery – The batter powers the RasberryPi and motors. I’m still trying to decide if I want a separate battery for the vacuum. However, I do see that there are lithium batteries that are specifically designed for cordless vacuums:(1) https://pethairvacuumcleaner.com/how-vacuums-work/cordless-vacuum-cleaner-batteries-motor-power-and-runtime (2) https://cleanhomeguide.co.uk/cordless-vacuum-lithium-ion-battery-guide/ . At a glance these cordless batteries are between $50 to $200 each.
Power Gage – will display the electricity power on available in an LED and on a computer UI.
RasberryPi – This is the computer that controls the robot. This needs to be flexible enough for multiple functions. There should be an easy way to replace features and then
Remote Control – the remote controls the robot through a computer. Also, there will be a wireless camera that will identify trash or obstacles. The operator will be near by but shouldn’t require following the robot around.
Wireless Camera – mounted on the front of the robot to capture images of trash or other obstacles. Will also be used to collect and label images of trash to hopefully enable the robot to work autonomously. There will still need to be a human to inspect the work and help the robot when it gets stuck but overtime hopefully that will be less and less necessary.
Vacuum – self explanatory. The vacuum will suck up the items of interest into the trash bin and trash bag.
Vacuum Tube – need to be large enough for most litter items. It will be larger than a typical home vacuum.
Trash Bin – self explanatory. It should hold the trash bag which can be taken on / off for emptying. A vacuum tube will be fed into the bin so the trash
Additional Design Thoughts:
Ideally the robot would know autonomously when it is near litter that it wants to pickup. I don’t want the robot to run the vacuum all the time because there is the possibility that it will pickup unnecessary things and possibly damage the environment that it’s working in. So, when it identifies trash it should turn the vacuum on for a specific period of time and then turn off then proceed through its route.
I think that we should be able to use the same base as the multi purpose robot frame. My goal is to make plug and play modules on the robot for many different services. Once I get one aspect of the business going this will help me to expand.
The feature to cut grass and other shrubs will probably need to be a second phase. I want to get the mobility features right before a start putting spinning blades on the device. We could also pick up leaves or grass as well as trash once we get the ‘mower’ feature integrated.
Initially to create the POC I want to use wood because it is cheaper. Later, I will machine metal parts once I get a better feel for what I’m doing. I will start by building the frame, motors and wheels which I’ll manage in a separate post.
The vacuum can be taken from a vacuum cleaner. Additionally, I could repurpose the tubes as well potentially. The tube needs to be fairly large for different litter types.
I could either ask family and friends for old or broken appliances. Additionally, I could deconstruct and store my own. I’ll need to figure out where to store these deconstructed parts and it would be nice to have a workshop to build and test my robot. However, that is a secondary concern to designing my POC and go to market strategy.
Path to Autonomy
As I gather video I’ll gather and label images that represent trash while remotely controlling the robot through a computer. Then I’ll feed the labeled images through a machine learning algorithm to train the robot.
I hypothesize that I could fence off an area that I want to clean and program to stop once the fence is reached. Then the robot just needs to up and down the surface area to vacuum up the trash and not the dirt and plants in the surrounding environment.
I’m also thinking to add sensors that could detect objects or walls that would impede the robots movement.
A human will still need to be involved to inspect the work done in each section and help the robot when it gets stuck. However, as the automation becomes better over time that should hopefully be required less and less.
Go to Market Strategy
The first iteration will prove the concept in a small contract with either a commercial real estate or state or city government to gain confidence for both them and myself that this can both efficiently / efficiently solve the littering problem and that it is profitable enough to be a sustaining business.
Record the work done on the project and post to YouTube. This will earn revenue and be a promotional mechanism.
Who makes decisions about how to manage litter in the city? Likely the Mayor will be an executive sponsor and will then redirect me to another official or government worker.
In the case of a property management company I’d imagine that it will depend on the size of the company and organizational structure.
I need to both network and research websites to source contacts.
I’ll use a web scraping tool and create my own SQLite database to save costs. I’ll create an article separately for more detail.
Once I start to build a network I may be able to facilitate other deals as well which could both expand my circle of influence and offer additional revenue opportunities.
I’ll create cold calling scripts and do some thought experiments for how the negotiations will go.
I’ll also create a presentation with the following detail:
The problem on a macro scale
The problem in a specific city
The cost of the alternatives
The cost for my services and bot
Pitch for a proof of concept to fix this in a specific area at a penetration for penetration pricing
Once I get agreement to a contract I’ll need to take the following operational steps as a matter of policy.
Get contractual signoff on the area for work.
Get signoff on the square feet or the time to spend
Take a picture of the area before
Take a picture of the after an after cleaning
The contract, images, operational, and cost metrics must be stored in a CRM system along with the account and contact information.
I could create a home grown business process management application or use something like Salesforce.
I’ll want to expand to other infrastructure like accounting and ERP Systems but that will come later once I get through a POC opportunity.
Asset maintenance will be another area of focus from BreakFix to Preventive Maintenance. It would be a good idea too to start capturing asset technical counters to start to figure out when to do Preventive Maintenance as the asset depredates from use. IoT is another option too which will allow me to automate parts orders or CRM record creation. I’ll create a plan and timing around these activities at some point after I get a contract for my first opportunity.
As I execute the clean up activities I’ll capture video to post on YouTube / my website and create website articles about operational procedure and lessons learned. That way I can generate advertising revenue and use it in promotional material for future contract opportunities.
What are my startup and operational costs?
Initial build of the robot – I’m tentatively forecasting that my POC costs will be around $1.5k. This could be more though if the motors or batteries end up being more expensive than I forecast.
Battery that powers wheel motors and vacuum
Robot Maintenance – I could see that I may need to replace parts if they break and/or upgrade parts of the robot to ones that are more durable or robust.
How to reduce energy costs?
Gas or electric seem the only options at this point
Electricity seems more aligned to the electronics I’m planning to use.
I could charge my device at a near by E-Vehicle station. However, I wonder if those costs may be a little high.
Solar is another option. I could either mount a solar panel on the robot or create a ‘solar station’ where I charge spare batteries. These batteries I would carry with me to a site and could swap on the robot while on the job.
Where to Go From There?
Once I start making money I can upgrade the current bot and create more as demand increases.
From there I’ll use the promotional material I’ve developed and iterate through my go to market strategy (hopefully more efficiently with each iteration).
I think I’ll have enough to figure out with the first phase of the POC but once I get everything down I’d like to add actuators onto the trash container to lift and dump the trash into a dumpster or the site that I will sort the trash.
I’d also like to build a separate bot that sorts litter into trash / recycling. I could even go further and start manufacturing with the reusable materials and resell.