Waterborne Autonomous Litter and Trash Extraction Robot(WALTER)

Problem: Lake Victoria is the world’s third largest lake, second largest freshwater lake

and largest lake in Africa. Unfortunately, the lake is dying, putting all of these lives at
risk. Plastic pollution is threatening to push it over the edge and lead to its complete
destruction. Lake Victoria is the source of the Nile, thus waste plastic may flow into the
Nile, affecting more communities along its path into the Mediterranean sea.
Problem Scale: Plastic waste piles up on beaches and the bed of the lake, further
damaging the remaining fish breeding areas that survived sand harvesting. This in turn
prevents fish from being able to successfully breed, further threatening already
declining populations. The fish also confuse some of the small plastic pieces for their
food, but are unable to digest them, leading to accumulation in their stomachs and,
eventually, death. The local communities’ animals also end up ingesting some of this
trash whilst drinking water or whilst grazing as some of the trash that collects on the
beach is blown further inland.
Microplastics also leach off of this waste, causing further damage. Communities and
towns depending on the lake for food and drinking water end up ingesting these
microplastics as they are unable to remove them from the water due to a lack of
treatment plants, even cities with water treatment plants, such as Kisumu, Kenya, lack
the proper technology to be able to filter out these microplastics. This is concerning as
microplastics have been found to be carcinogens and some studies have linked them
to infertility.

The contamination of the food chain due to the use of polluted water for agricultural
purposes leads to introduction of toxins in food which affects the overall health of the
consumers - An estimated 600 million – almost 1 in 10 people in the world – fall ill after
eating contaminated food and 420 000 die every year, resulting in the loss of 33 million
healthy life years according to the World Health Organization(WHO).
Plastic pollution has rapidly risen to become one of the world’s most pressing
problems. The problem is only growing and the communities suffering from it are rarely
ever responsible yet face the worst effect as it is their livelihoods that are threatened
by the loss of these ecosystems.
Number of people affected: Lake Victoria supports over 4 million people in Kenya,
Uganda and Tanzania. Many of them rely on it for their livelilihoods such as fishing,
agriculture and water.
Relation to the Solution: While clean-up campaigns can collect land based trash, it is
more difficult addressing pollution in water bodies. Floating trash is easier to collect
using boats than underwater trash which requires laborious efforts from a large group
of people. To add on, the manual process of using means such as diving to collect
litter can prove to be quite time consuming, hence by implementing our autonomous
vehicle we aim to break the monotony and restore confidence in the technology of the
future.

What it is: WALTER is an autonomous vessel that uses AI to collect non-

biodegradable waste and data on waste from polluted water bodies. It is equipped with
mechanisms such as: a robotic claw to optimize collection; a sonar (Sound Navigation
Ranging); and a cavity that is separated to collect different forms of litter to ensure
quick and safe disposal.
What it does: WALTER collects non-biodegradable waste, mainly plastics, from the
surface of and, in future, on the bed of polluted water bodies.
How it works: The process of collecting rubbish within the vessel is quite simple.
Firstly, the drone’s brain, a single board computer, receives input from the

camera then uses a model of the YOLO algorithm that we have trained to identify the
trash. The drone then pinpoints the location of the trash relative to itself and begins
tracking it. It then proceeds to collect the material through one of two methods: The
first is approaching the rubbish and locking it up in the cavity. The second uses a
robotic claw to collect litter in hard to reach areas or in areas that may cause damage
to the drone if it proceeds, such as those with water hyacinth. Upon collecting the
trash, the drone logs it within its database, This procedure continues until the
maximum storage has been reached whereby it navigates to a dock specified by the
operator where litter can be safely be removed. It then returns to its life changing task
of waste collection.
Technology: The technological aspect of this drone makes it stand out. The
coordination to ensure optimum function is powered by hi-tec systems that have been
tried and tested by industry leaders such as Google, Toyota and Intel. An example of
such a system is OpenCV, a program which assists the components in-charge of
visual recognition to maximise accuracy when collecting the rubbish. The YOLO
algorithm and the addition of SONAR is essential in avoiding damage to the drone and
the aquatic ecosystem itself as it aids in detecting and determining the distance and
direction of underwater objects. This in turn helps in maintaining accuracy and
efficiency in poor visibility conditions, whilst keeping environmental harm to a minimal
level. Moreover, WALTER runs on solar energy which not only reduces the cost of
operation and keeps it as eco-friendly as possible, but also means that it is able to
operate for long durations without requiring to recharge, thus allowing higher
operational efficiency to be achieved. Continuing with the theme of sustainability, we
are constructing the most of the drone out of recyclable materials to ensure that
replaced parts can be kept out of the very waterbodies we are cleaning. Finally, we are
making the body adaptable such that should its use as a trash collector cease to be
required in a particular water body, it can easily be modified to perform water
monitoring tasks, thus supporting waste management.

Target Population: We are targeting the local communities spread out across the

shores of lake Victoria, starting with Usoma and Kogony communities in Kisumu. Many
of them rely on the lake not only for food and water for domestic use, but also for their
livelihoods.
Who they are and ways in which they are underserved: Fishermen depend almost
solely on the lake to provide for their families. The slow but sure degradation of the
lake has led to rapidly falling fish populations, forcing many of the fishermen to go out

earlier and stay out on the lake longer whilst still getting a significantly lower catch than
a few decades ago. This leaves them with less time to spend with their families. The
lower fish catch also means that they find it much harder to earn enough money to
support their families' most basic needs, such as healthcare, education, food and
clothing. This in turn leads to more children being forced to drop out of school with
others helping their parents in their work to try to earn more money. Some of the youth
engage in destructive activities as a result. It also causes the destructive cycle to
repeat as without proper education, these children have drastically reduced access to
opportunities, causing them to undertake the same work as their parents regardless of
what they aspire to do. Women fish retailers are also negatively affected as they would
use these now plastic clogged beaches to purchase fish from fishermen, descale
them, meet with other community members and even sell the fish. However these
beaches are too unsanitary to be used. Mother’s also used these beaches to wash
dishes and clothes and fetch water for drinking, but now are unable to do so and
forced to travel longer distances in search of cleaner places and water. The most
devastating part is that majority of the very waste that litters their beaches isn’t
generated by them but flows in through city drainages and is brought in by rivers. The
proof of this is the presence of brands that either aren’t sold locally, or too expensive to
be consumed at such scale by community members.
How the solution will address their needs: WALTER will intercept and extract this
trash before it reaches the beaches and settles on breeding sites. This will lead to less
trash on beaches, making the beaches cleaner and safer for human use. Less plastic
on breeding sites will allow more fish to breed and help to stabilise the fish populations
and eventually contribute to their recovery. This will in turn help the fishermen earn
more money to provide for their families and give the community back for their
beaches. Women fish retailers will get more fish and work on cleaner beaches. The
community will have cleaner water and environment. The communities will be able to
sell the recyclable plastic that the drone collects to recycling centres within their towns
earning incomes to better their lives.

Appropriateness of the Team & Team Lead Community Representation: Our team

lead, Jeremy Muchilwa, is part of the Usoma Community, which is one of many

suffering from the effects of plastic pollution and the lakes death. Being raised on the
shores of the lake has meant that he has been one of the millions who depend on the
lake witnessing its slow but sure degradation. Seeing the lakes problems and being
able to talk to elders who have lived on its shores for decades also provides a better
understanding of the problems it faces and how they procedurally came about. Our
head engineer often visits his hometown of Chittagong, Bangladesh. Every few
months, Lake Boga which is next to his house dries up by a few metres, revealing
heaps of trash and rubbish. He is astonished by the ever increasing amount of litter.
During his visits to Lake Victoria and the Usoma Community during the holidays, he
has experienced all the heaps of trash suffocating Lake Victoria. He has correlated
and learnt from both experiences, making it easier to create effective designs. We
have lived next to Nairobi River, where we have experienced the problem of plastic
population. The river become an open sewer, from the release of non-biodegradable
and organic waste from industries, commercial centres, and residential areas. As a
result, the water is unsafe to use and there is increased heavy flooding in Nairobi
during heavy rains. Furthermore, the river has long lost its fauna that included several
fish species from plastic and heavy metals affecting feeding and breeding grounds.
Despite having organised many clean up campaigns along the lake in Kisumu, we
have not been able to reach the waste in the lake. These experiences has inspired us
to work on a drone to collect waste in water bodies.
Community Input, Ideas and Agenda in the Solution: The development of the
drone is heavily influenced by input and the needs of the community. The community
has expressed frustration with plastic and diapers polluting and clogging their water
sources and want it removed. They can no longer drink the water, bath or wash
clothes. The fishermen has complained about catching more plastic and diapers than
fish. They want a clean lake. Removing this plastic serves a key community need.
Following their input, the drones navigation has been designed to keep a certain
distance from heavily populated beaches to assure that the communities activities are
not disrupted and also to prevent possible theft. It will also avoid fishermen’s boats and
nets and they voiced concerns over it chasing away fish whilst operating. Essentially,
the drone is designed to integrate itself into the community's activities without being
disruptive. The varying needs of different communities also motivated us to make it
modular so that different communities can have different attachments put in. For
instance, some may not require the claws and some of the sensors, thus can save
money by omitting it.

So far, we have constructed and tested 3 prototypes of WALTER in environments ranging from ponds to Lake Victoria. The latest model we have developed contains the majority of the core features, namely: plastic detection, collection, tracking and rudimentary data logging, changeable buoyancy and manual control of the claw. We have run some tests in the presence of the community to allow them to see the capabilities of the drone and comment on necessary improvements.

Beyond gaining access to funding in order to continue our research and development

process, we are applying to Solve for an additional four key reasons that we believe
will be consequential to the development of our project.
Financial: Currently, the development of our project has been largely strained by
insufficient funds. This has slowed down the procurement of equipment necessary to
allow W.A.L.T.E.R. to perform select functions more effectively. This has led to either
slow or less efficient operation of systems such as plastic detection (due to the
mainboard having limited power), propulsion and the robotic claw. Mentors will also
provide insights on effectively pitching our product to potential funders.
Marketing: The first concerns access to market education and mentorship. As we
develop our project with the intent of it becoming a product and competitor in the waste
management market, we are seeking mentors and learning opportunities that will help
us market W.A.L.T.E.R. more effectively as a product to the relevant consumers. We
will gain additional perspectives on what key or new features should be focused on in
order to gain an edge over market competitors and be more impactful whilst still
keeping it financially viable for the market.
Legal: The former point connects with our second reason, which is gaining legal
mentorship, which, would help us gain a better understanding of the intellectual
property and patenting space. This would not only help us better protect our own
innovations and guide us through the application process for patents but also help us

be more cautious about what patent-protected areas we may be stepping into with our
project’s development.
Technical: Solve provides the opportunity to connect with industry leaders. Currently,
our network of professionals in fields related to our project is very limited. This has
made it harder to validate or receive consultation on our approaches and ideas. Solve
could provide us the opportunity to receive technical insights into some of the core
technologies that the drone currently is and will implement, as well as advice on how to
begin our entrepreneurial journey with our project. Lastly, we believe that the network
provided by MIT Solve will be an environment impactful in the bouncing of ideas that
will positively affect our project’s development and could possibly help in the
recruitment of highly experienced and talented individuals to grow our project. This will
address our need for additional skills, labour, and insights as our project develops in
complexity.

Different Solution Approach to the Problem: The main method used for collecting

surface level plastic is to hire boats and go out into the water body with poles, bags
and nets to collect the plastic. As for plastic on the bed of water bodies, the only two
options are either to hire divers to go and fetch the plastic, or to perform bottom
trawling. All of these methods are expensive, especially diving and trawling, and
require a lot of time and human involvement. These reasons mean that clean-ups are
often restricted to land due to the parties carrying them out not finding it financially
viable to hold multiple offshore clean-ups. This limits the scale of impact clean-ups
have since waste from the water is washed onto the beach, returning it to its initial
state. WALTER helps to alleviate the majority of the cost, time and human involvement
necessary to carry out these clean-ups through its autonomous operation and use of
renewable energy. One or two operators to take out the trash collected by WALTER. It
has an opt-in data storage and management service that supports intelligent waste
management.
Catalysing Broader Positive Impacts: WALTER can be used in many other marine
bodies – not just Lake Victoria. This will in turn allow for more effective and efficient
clean-ups in lakes, rivers, oceans and seas. WALTER will inspire others, especially
youth, to capitalise on technology to address the plastic menace. They will be able to
design and deploy even better drones following the inspiration from WALTER.

Changing the Market Place/Landscape: WALTER is able to operate daily, leading to
many more tonnes of trash being collected. This will change current clean-up practices
from occasional campaigns to long term efforts that have greater impact. It will support
the greater use of technology, drones and AI in waste management and restoration of
degraded water bodies and systems. Furthermore, WALTER also collects data on the
trash it is collecting, such as GPS location, brand and, if identifiable, type. This data
can then be used to aid in finding the possible origin of some of the trash (as certain
brands are only found in certain regions and in extreme circumstances travel across
nations). This data is crucial for organisations such as NGOS working to curb pollution
as it not only allows them to track the origin of the waste and whether progress is
being made. WALTER will also target underwater trash which will help to get rid of the
carpet of plastic forming on beds of water bodies such as Lake Victoria's, helping to
preserve key breeding sites and drastically increasing the scope that clean-up’s have.

Plastic pollution in Lake Victoria and other water bodies is damaging the environment,

killing biodiversity and affecting livelihoods. It poses huge threats to the health of
humans and biodiversity as a whole. It undermines livelihood activities such as fishing
and tourism as well as business activities such as water treatment, power generation
and transport. Plastic pollution is an eyesore that undermines the standards of living of
neighbouring communities. While land based plastic waste is easier to collect, recycle
and dispose of, water based plastic is often out of reach of conventional clean-up
approaches and campaigns. It demand expensive approached such as trawling and
diving to remove forcing many environmental actors to ignore it altogether.
By leveraging technologies such as drones and Artificial Intelligence, and capacitating
communities and organisations on their use, more cost effective, efficient and
sustainable approaches will be used to remove plastic from polluted water systems.
Technology will provide more information on the extent and location of plastic pollution
in water bodies and support more effective research and waste management
practices. This will result in reduced plastic waste in water bodies thus reducing the
prevalence of micro-plastics while enabling biodiversity to flourish. The cleaner
environment will increase fish yields thus supporting community livelihoods as well as
the availability of clean water for drinking, domestic chores and agriculture. Business
activities such as tourism, fishing, power generation and water treatment will thrive

supporting economic development and standards of living. The environment and its
rich biodiversity will be preserved for current and future generations.

Sustainable Development Goals (SDGs): The aims that guide the development of

the drone closely align with the United Nations SDG 6 (Clean Water and Sanitation)
and 14 (Life Below Water). More specifically, the drone aims to tackle 6.3 (the
improvement of water quality by reducing pollution, eliminating dumping and
minimising release of hazardous chemicals and materials, halving the proportion of
untreated wastewater and substantially increasing safe reuse globally) in Lake Victoria
and other water bodies the drone shall be deployed in through the surface and
underwater level plastic collection, 6.b (Support and strengthen the participation of
local communities in improving water and sanitation management) through the
provision of drones to local communities and non-governmental organisation for use in
the water bodies they plan to implement it in, and 14.1 (By 2025, prevent and
significantly reduce marine pollution of all kinds, in particular from land-based
activities, including marine debris and nutrient pollution).
As we develop our drone in line with the SDGs. These are the key goals that we aim to
achieve are:
Impact Goal 1: Reduced plastic pollution on the beaches and waters of Lake
Victoria.
It is our aim to significantly decrease plastic and macro-plastic waste found within
Lake Victoria ( as well as other water bodies in which the drone shall be deployed in
future). To track our progress, we will use three main indicators.

● Indicator 1: Changes in the amount of plastic waste collected along the beach.
This shall be tracked using clean-up data, which often has data that includes
the number and type of waste collected.
● Indicator 2: Changes in the amount of estimated plastic waste on the bed of
Lake Victoria around Kisumu. This will be identifiable through data collected by
trawls and other scientific studies carried out by governmental and research
bodies present in the given waterbody, and also through observing maps being
actively created by the drone’s mapping system.
● Indicator 3: % of plastic captured in fish net catches: We shall rely on reports
from local fishermen, who often fish and find plastic in the lake, to identify any
decrease in plastic, and increase in fish yields.
Impact Goal 2: Availability of data on plastic waste found and collected in Lake
Victoria. This involves making a database for plastics identified through collection.
This will be highly important as currently no database exists categorizing the plastic
waste found in the lake, which not only makes it more difficult to categorize and
identify the major contributors towards pollution in the lake, but also helps in the
development of AI models that can be trained in the further identification, collection,
and processing of such waste in future projects. This impact goal has three main
indicators of success:
● The existence of a data pool on plastic waste in Lake Victoria
● Amount of data collected on waste pollution in Lake Victoria

Waste Plastic Recognition: To begin with, the main system for plastic recognition is

powered by the You Only Look Once or YOLO algorithm. The algorithm provides a
good balance between speed and accuracy of detection.
Robotic Claw: The use of a robotic claw provides the drone with the ability to collect
trash underwater. It also provide the ability to extract trash that is trapped in nets or
stuck in hyacinth without the risk of getting the drone stuck or damaged. The claw is
being adapted to work with the DexNet algorithm from UC Berkeley. Work is currently

being done to try to get the algorithm and claw to both work favourably despite the
rapidly altering nature of the environment they are expected to function in. To assist
both of these systems, a depth camera will be used in conjunction with a SONAR to
allow the drone to be able to interpret the environment around it, preventing unwanted
collisions, and making it easier for the claw to collect the plastic both above and
underwater.
Navigation: Moreover, the drone uses GPS to navigate to preset docking locations to
drop off the trash and or come back for collection at the end of the day. Additionally,
the operator will be able to monitor the operation of the drone through either a website
or mobile app implementing the Google Maps API in order, ensuring that the operator
will always have access to the vessel to be able to respond to any problems or
manually take control should the need arise. These platforms will also allow the
operator to view the data of what the drone has and continues to collect.
Micro-plastic Collection: We are working on utilising hydrogels to absorb the
microplastics and trap them within their network as the drones move. These
microplastics can then be broken down using plastic eating enzymes.
Mapping: Finally, we are utilising navigation algorithms such as Simultaneous
Localisation And Mapping(SLAM), to aid in construction of 3D maps of the areas the
drones are working to give operators a better view of the underwater environment, and
at least provide other drones in the fleet with a basic idea of what to expect in the
environments, despite the dynamic nature of the environment in which they operate.

We began work on WALTER in January 2023. Thus we have been working on it for 1 year and 3 months.

Diversity: Our team is composed of people from multiple different ethnic backgrounds
with different cultural and religious beliefs. The core team consists of people ethnically
from Kenya, Somalia, Sri Lanka, Pakistan, Bangladesh and Oman. Despite having
different religious and cultural beliefs, we have and continue to work effectively
between our various departments to develop the drone.
Inclusion: Furthermore, we solely bring members based on what they are able to
contribute to the drones and also on them proving that they are able to effectively work
with the other members without judging or discriminating anyone on their beliefs or
certain actions.
Equity: Additionally, decisions are made as a group whilst accounting for everyone's
needs and ideas. We do this not only to make sure everyone is satisfied with our
trajectory, but also because we have found it to be much more ideal than relying on
one person to do it as it provides more perspectives all of which are influenced by
varying upbringings.

Goals: We are currently recruiting more members to the team to aid as we expand
operations and continue to add on the promised features as we improve prototypes.
We are recruiting members from various different areas, ranging from separate
schools to, for roles that can be carried out online, different cities. Casting a wider net
as to where we get people also helps us locate people with better skills. Even though
they may not be directly involved in things like programming, we involve community
members as advisers to the design and functions of the drone. We also involve young
people and women from other communities through mentoring them and helping them
explore their interests in technology through learning about how some of WALTER’s
systems work and being allowed to participate in tests and go on to hold their own
presentations on the drone.

Key Customers and Beneficiaries: The target demographic for our pioneering

solution spans a wide spectrum of stakeholders, ranging from government agencies
and research organizations to NGOs and environmentally conscious businesses. The
final beneficiaries will include communities who benefit from cleaner water and
environments, higher fish catches and thus better livelihoods. This particular audience
will not pay for the product as opposed to the former who will actually purchase the
services. Central to our operational paradigm is a steadfast dedication to community

empowerment. By offering comprehensive training initiatives and fostering
employment opportunities, we aim to cultivate a proficient workforce capable of
effectively managing our drone fleet.
Product and Service Need: The customers face challenges with polluted waters,
including plastic pollution that interfered with business provision. This includes fishing,
tourism, water treatment, agriculture, health and environmental conservation.
Restaurants, hotels and other tourism facilities lose customers due to polluted waters
bordering their facilities. It also undermines their competitive edge. Power and water
treatment companies have to spend money cleaning out plastic as it interferes their
equipment. It undermines the quality of water sold to customers due to micro plastics.
Government bodies such as Kisumu city are mandated to provide services and
safeguard the environment. Others such as the Fisheries Department and Kenya
Wildlife Services are mandated to conserve the environment and its biodiversity.
Kenya Marine and Fisheries Institute supports research. The products are therefore
needed to improve the business environment as well as meet the organisation
mandate. Designed to address the burgeoning challenge of non-biodegradable waste
accumulation in polluted water bodies, our autonomous drone represents a paradigm
shift in environmental stewardship. By catering to the diverse needs and financial
capacities of our clientele, we aim to democratize access to cutting-edge technology
while effecting tangible, positive change on a global scale.
Product and Services: These will include (1) Drone product sales, (2) Providing users
access to drone data (3) Drone product repairs and maintenance services (4)
component sales (5) Trash collection services, and leasing
Our product and revenue model is multifaceted, encompassing diverse income
streams meticulously tailored to meet the needs of our discerning clientele. Foremost
among these is the sale of our cutting-edge autonomous drones, augmented by
supplementary revenue derived from the provision of maintenance, repair, and
enhancement services. Additionally, we offer a bespoke data management solution,
affording clients seamless access to actionable insights gleaned from the
comprehensive data sets collected by our drones. This value-added service not only
enhances operational efficiency but also underscores our unwavering commitment to
customer-centric innovation.
Product and Service Provision: Drones will be sold to interested customers. In
addition, we will provide drone renting options and offer data collection and
management services, with proceeds reinvested into research, operational costs, and
supporting related projects. Drone repair services will also be offered. Repair fees are
based on the cost of repair or, if irreparable, the price of a new drone. The fees will be
affordable. This comprehensive approach underscores our commitment to
sustainability and innovation.

Initial Development: We currently have a drone that has some of the features and

provide proof of concept. The bulk of funding that has supported the drone
development has come from our parents and our own funds. In addition, through the
Captain Planet Foundation, we have received $1000 USD for research and
development through their fundraising platform, which we set up and plan to continue
using as we continue early stage development. The limited funding for done
development has meant purchasing less appropriate components in the desire to
prove the concept can work.
Future Development: In order to carry out larger scale research and development,
which includes updates to most of the drone’s equipment, and developing new
features (which may require the procurement of new equipment), we will rely on grants
and fundraising to meet the costs. This will be easier with the simple done that
provides proof of concept.
Product and Service Sales: As the drone is developed and reaches further stages in
development that should allow it to become a viable product, the main streams of
revenue which it should rely on include the selling of drone models as well as parts of
it to community, research and non-governmental organisations for use in their target
water bodies. Other than for maintenance, parts to upgrade and or expand the
capabilities of the drone will also be sold. Furthermore, since the drone will be actively
collecting data on the aquatic environments it shall work in, we plan to provide
summarised analytical reports of pollution collected within the lake on an interactive
and easy to interpret an export from platform as a means to aid customers in research

where it shall be relevant. We will also provide a data storage service for the drone’s
users to help individuals or organisations that may not have the equipment to store all
the data that the drone collects. Lastly, for those who may not be interested in buying
the drone but would still require its services, we will offer a renting program that will
base cost on the number of drones rented, and the duration of use.
In terms of financing, we adopt a pragmatic approach, leveraging grants and
philanthropic contributions to fuel our initial developmental endeavors. As we transition
into operational viability and revenue generation, sustained financial solvency will be
ensured through judicious fiscal management and the prudent allocation of incoming
funds. This adaptive financing strategy underscores our steadfast commitment to long-
term sustainability and underscores our unwavering dedication to realizing our vision
of a cleaner, more sustainable future.